Appendix G Solutions to Exercises
1 Integration
1.1 Definition of the Integral
1.1.8 Exercises
1.1.8.2.
Solution.
-
Solution 1: One naive way to solve this is to simply use the same method as Question 1.The rectangle on the left has area \(3 \times 2.25 = 6.75\) square units, and encompasses the entire shaded region. The rectangle on the right has area \(3 \times 0.25 = 0.75\) square units, and is entirely contained inside the blue-shaded region. So, the area of the blue-shaded region is between 0.75 and 6.75 square units.This is a legitimate approximation, but we can easily do much better. The shape of this graph suggests that using the areas of three rectangles would be a natural way to improve our estimate.
-
Solution 2: Let’s use these rectangles instead:In the left picture, the red area is \((1 \times 1.25)+(1 \times 2.25)+(1 \times 0.75)=4.25\) square units. In the right picture, the red area is \((1 \times 0.75)+(1 \times 1.75)+(1 \times 0.25)=2.75\) square units. So, the blue shaded area is between 2.75 and 4.25 square units.
1.1.8.3.
Solution.
-
Try 1: First, we can try by using a single rectangle as an overestimate, and a single rectangle as an underestimate.The area under the curve is less than the area of the rectangle on the left (\(2 \times \frac{1}{2}=1\)) and greater than the area of the rectangle on the right (\(2 \times \frac{1}{8}=\frac{1}{4}\)). So, the area is in the range \(\left(\frac{1}{4},1\right)\text{.}\) Unfortunately, this range is too big--we need our range to have length at most 0.2. So, we refine our approximation by using more rectangles.
-
Try 2: Let’s try using two rectangles each for the upper and lower bounds.The rectangles in the left picture have area \(\left(1 \times \frac{1}{2}\right)+\left(1 \times \frac{1}{4}\right)=\frac{3}{4}\text{,}\) and the rectangles in the right picture have area \(\left(1 \times \frac{1}{4}\right)+\left(1 \times \frac{1}{8}\right)=\frac{3}{8}\text{.}\) So, the area under the curve is in the interval \(\left(\frac{3}{8},\frac{3}{4}\right)\text{.}\) The length of this interval is \(\frac{3}{8}\text{,}\) and \(\frac{3}{8} \gt \frac{3}{15}=\frac{1}{5}=0.2\text{.}\) (Indeed, \(\frac{3}{8}=0.375 \gt 0.2\text{.}\)) Since the length of our interval is still bigger than 0.2, we need even more rectangles.
-
Try 3: Let’s go ahead and try four rectangles each for the upper and lower estimates.The area of the rectangles on the left is:\begin{align*} \left(\frac{1}{2}\times \frac{1}{2}\right)+ \left(\frac{1}{2}\times \frac{1}{2\sqrt{2}}\right)+ \left(\frac{1}{2}\times \frac{1}{4}\right)+ \left(\frac{1}{2}\times \frac{1}{4\sqrt{2}}\right) &\\ = \frac{3}{8}\left[1+\frac{1}{\sqrt{2}}\right],& \end{align*}and the area of the rectangles on the right is:\begin{align*} \left(\frac{1}{2}\times \frac{1}{2\sqrt{2}}\right)+ \left(\frac{1}{2}\times \frac{1}{4}\right)+ \left(\frac{1}{2}\times \frac{1}{4\sqrt{2}}\right)+ \left(\frac{1}{2}\times \frac{1}{8}\right) &\\ = \frac{3}{8}\left[\frac{1}{2}+\frac{1}{\sqrt{2}}\right].& \end{align*}So, the area under the curve is in the interval \(\left( \frac{3}{8}\left[\frac{1}{2}+\frac{1}{\sqrt{2}}\right], \frac{3}{8}\left[1+\frac{1}{\sqrt{2}}\right]\right)\text{.}\) The length of this interval is \(\frac{3}{16}\text{,}\) and \(\frac{3}{16} \lt \frac{3}{15}=\frac{1}{5}=0.2\text{,}\) as desired. (Indeed, \(\frac{3}{16}=0.1875 \lt 0.2\text{.}\))Note, if we choose any value in the interval \(\left( \frac{3}{8}\left[\frac{1}{2}+\frac{1}{\sqrt{2}}\right], \frac{3}{8}\left[1+\frac{1}{\sqrt{2}}\right]\right)\) as an approximation for the area under the curve, our error is no more than 0.2.
1.1.8.4.
Solution.
1.1.8.5.
Solution.
1.1.8.6.
Solution.
- Two possible answers are \(\displaystyle\sum_{i=3}^7 i\) and \(\displaystyle\sum_{i=1}^5 (i+2)\text{.}\) The first has simpler terms (\(i\) versus \(i+2\)), while the second has simpler indices (we often like to start at \(i=1\)). Neither is objectively better than the other, but depending on your purposes you might find one more useful.
- The terms of this sum are each double the terms of the sum from part (a), so two possible answers are \(\displaystyle\sum_{i=3}^7 2i\) and \(\displaystyle\sum_{i=1}^5 (2i+4)\text{.}\) We often want to write a sum that involves even numbers: it will be useful for you to remember that the term \(2i\) (with index \(i\)) generates evens.
-
The terms of this sum are each one more than the terms of the sum from part (b), so two possible answers are \(\displaystyle\sum_{i=3}^7 (2i+1)\) and \(\displaystyle\sum_{i=1}^5 (2i+5)\text{.}\)In the last part, we used the expression \(2i\) to generate even numbers; \(2i+1\) will generate odds. So will the index \(2i+5\text{,}\) and indeed, \(2i+k\) for any odd number \(k\text{.}\) The choice of what you add will depend on the limits of \(i\text{.}\)
-
This sum adds up the odd numbers from 1 to 15. From Part (c), we know that the formula \(2i+1\) is a simple way of generating odd numbers. Since our first term should be 1 and our last term should be 15, if we use \(\sum (2i+1)\text{,}\) then \(i\) should run from \(0\) to \(7\text{.}\) So, one way of expressing our sum in sigma notation is \(\displaystyle\sum_{i=0}^7 (2i+1)\text{.}\)Sometimes we like our sum to start at \(i=1\) instead of \(i=0\text{.}\) If this is our desire, we can use \(2i-1\) as our terms, and let \(i\) run from 1 to 8. This gives us another way of expressing our sum: \(\displaystyle\sum_{i=1}^8 (2i-1)\text{.}\)
1.1.8.7.
Solution.
- The denominators are successive powers of three, so one way of writing this is \(\displaystyle\sum_{i=1}^4 \frac{1}{3^i}\text{.}\) Equivalently, the terms we’re adding are powers of \(1/3\text{,}\) so we can also write \(\displaystyle\sum_{i=1}^4 \left(\frac{1}{3}\right)^i\text{.}\)
- This sum is obtained from the sum in (a) by multiplying each term by two, so we can write \(\displaystyle\sum_{i=1}^4 \frac{2}{3^i}\) or \(\displaystyle\sum_{i=1}^4 2\left(\frac{1}{3}\right)^i\text{.}\)
- The difference between this sum and the previous sum is its alternating sign, minus-plus-minus-plus. This behaviour appears when we raise a negative number to successive powers. We can multiply each term by \((-1)^i\text{,}\) or we can slip a negative into the number that is already raised to the power \(i\text{:}\) \(\displaystyle\sum_{i=1}^4(-1)^i \frac{2}{3^i}\,,\) or \(\displaystyle\sum_{i=1}^4 \frac{2}{(-3)^i}\text{.}\)
-
This sum is the negative of the sum in part (c), so we can simply multiply each term by negative one: \(\displaystyle\sum_{i=1}^4(-1)^{i+1} \frac{2}{3^i}\) or \(\displaystyle\sum_{i=1}^4 -\frac{2}{(-3)^i}\text{.}\)Be careful with the second form: a common mistake is to think that \(-\dfrac{2}{(-3)^i} = \dfrac{2}{3^i}\text{,}\) but these are not the same.
1.1.8.8.
Solution.
-
If we re-write the second term as \(\frac{3}{9}\) instead of \(\frac{1}{3}\text{,}\) our sum becomes:\begin{equation*} \frac{1}{3}+\frac{3}{9}+\frac{5}{27}+\frac{7}{81}+\frac{9}{243} \end{equation*}The numerators are the first five odd numbers, and the denominators are the first five positive powers of 3. We learned how to generate odd numbers in Question 6, and we learned how to generate powers of three in Question 7. Combining these, we can write our sum as \(\displaystyle\sum_{i=1}^5 \frac{2i-1}{3^i}\text{.}\)
- The denominators of these terms differ from the denominators of part (a) by precisely two, while the numerators are simply 1. So, we can modify our previous answer: \(\displaystyle\sum_{i=1}^5 \frac{1}{3^i+2}\text{.}\)
- Let’s re-write the sum to make the pattern clearer.
1.1.8.9.
Solution.
- Using Theorem 1.1.6, part (a) with \(a=1\text{,}\) \(r=\frac{3}{5}\) and \(n=100\text{:}\)\begin{equation*} \sum_{i=0}^{100} \left(\dfrac{3}{5}\right)^i = \dfrac{1-\left(\frac{3}{5}\right)^{101}}{1-\frac{3}{5}} = \dfrac{5}{2}\left[1-\left(\frac{3}{5}\right)^{101}\right] \end{equation*}
- We want to use Theorem 1.1.6, part (a) again, but our sum doesn’t start at \(\left(\frac{3}{5}\right)^0=1\text{.}\) We have two options: factor out the leading term, or use the difference of two sums that start where we want them to.
- Solution 1: In this solution, we’ll make our sum start at 1 by factoring out the leading term. We wrote our work out the long way (expanding the sigma into “dot-dot-dot” notation) for clarity, but it’s faster to do the algebra in sigma notation all the way through.\begin{align*} \displaystyle\sum_{i=50}^{100} \left(\dfrac{3}{5}\right)^i&= \left(\dfrac{3}{5}\right)^{50}+ \left(\dfrac{3}{5}\right)^{51}+ \left(\dfrac{3}{5}\right)^{52}+\cdots+ \left(\dfrac{3}{5}\right)^{100}\\ &= \left(\dfrac{3}{5}\right)^{50}\left[1+ \left(\dfrac{3}{5}\right)+ \left(\dfrac{3}{5}\right)^{2}+\cdots+ \left(\dfrac{3}{5}\right)^{50}\right]\\ &= \left(\dfrac{3}{5}\right)^{50}\dfrac{1-\left(\frac{3}{5}\right)^{51}}{1-\frac{3}{5}}\\ &=\dfrac{5}{2}\left(\dfrac{3}{5}\right)^{50}\left[1-\left(\frac{3}{5}\right)^{51}\right]. \end{align*}
- Solution 2: In this solution, we write our given expression as the difference of two sums, both starting at \(i=0\text{.}\)\begin{align*} \displaystyle\sum_{i=50}^{100} \left(\dfrac{3}{5}\right)^i&= \displaystyle\sum_{i=0}^{100} \left(\dfrac{3}{5}\right)^i- \displaystyle\sum_{i=0}^{49} \left(\dfrac{3}{5}\right)^i\\ &=\dfrac{1-\left(\frac{3}{5}\right)^{101}}{1-\frac{3}{5}} - \dfrac{1-\left(\frac{3}{5}\right)^{50}}{1-\frac{3}{5}}\\ &=\dfrac{5}{2}\left[\left(\frac{3}{5}\right)^{50}-\left(\frac{3}{5}\right)^{101}\right]\\ &=\dfrac{5}{2}\left(\dfrac{3}{5}\right)^{50}\left[1-\left(\frac{3}{5}\right)^{51}\right]. \end{align*}
- Before we can use the equations in Theorem 1.1.6, we’ll need to do a little simplification.\begin{align*} \displaystyle\sum_{i=1}^{10} \left(i^2-3i+5\right)&= \displaystyle\sum_{i=1}^{10} i^2 +\displaystyle\sum_{i=1}^{10} -3i +\displaystyle\sum_{i=1}^{10}5\\ &= \displaystyle\sum_{i=1}^{10} i^2 -3\displaystyle\sum_{i=1}^{10} i +5\displaystyle\sum_{i=1}^{10}1\\ &= \frac{1}{6}(10)(11)(21) -3\left(\frac{1}{2}(10\cdot 11)\right) +5\cdot 10\\ &=270 \end{align*}
- As in part (c), we’ll simplify first. The first part (shown here in red) is a geometric sum, but it does not start at \(1=\left(\frac{1}{e}\right)^0\text{.}\)\begin{align*} \displaystyle\sum_{n=1}^{b}\left[\textcolor{red}{ \left(\frac{1}{e}\right)^n} +\,\textcolor{blue}{en^3}\right]&= \textcolor{red}{\displaystyle\sum_{n=1}^{b} \left(\frac{1}{e}\right)^n} + \textcolor{blue}{ \displaystyle\sum_{n=1}^{b}en^3}\\ &=\textcolor{red}{\displaystyle\sum_{n=0}^{b} \left(\frac{1}{e}\right)^{n}-1} + \textcolor{blue}{e\displaystyle\sum_{n=1}^{b}n^3}\\ &=\textcolor{red}{\dfrac{1-\left(\frac{1}{e}\right)^{b+1}}{1-\frac{1}{e}}-1} + \textcolor{blue}{e\left[\frac{1}{2}b(b+1)\right]^2}\\ &=\textcolor{red}{\dfrac{\frac{1}{e}-\left(\frac{1}{e}\right)^{b+1}}{1-\frac{1}{e}}} + \textcolor{blue}{e\left[\frac{1}{2}b(b+1)\right]^2}\\ &=\textcolor{red}{\dfrac{1-\left(\frac{1}{e}\right)^b}{e-1}}+\textcolor{blue}{\frac{e}{4}\left[b(b+1)\right]^2} \end{align*}
1.1.8.10.
Solution.
- The two pieces are very similar, which we can see by changing the index, or expanding them out:\begin{align*} &\displaystyle\sum_{i=50}^{100} (i-50)+\displaystyle\sum_{i=0}^{50} i\\ &= \left(0+1+2+\cdots + 50\right)+\left(0+1+2+\cdots + 50\right)\\ &=\left(1+2+\cdots + 50\right)+\left(1+2+\cdots + 50\right)\\ &=2\left(1+2+\cdots + 50\right)\\ &=2\sum_{i=1}^{50} i\\ &= 2\left(\frac{50\cdot 51}{2}\right)=50\cdot 51=2550 \end{align*}
- If we expand \((i-5)^3 = i^3-15i^2+75i-125\text{,}\) we can break the sum into four parts, and evaluate each separately. However, it is much simpler to change the index and make the term \((i-5)^3\) into \(i^3\text{.}\)\begin{align*} \displaystyle\sum_{i=10}^{100} \left(i-5\right)^3&= 5^3+6^3+7^3+\cdots +95^3\\ \end{align*}
We have a formula to evaluate the sum of cubes if they start at \(1\text{,}\) so we turn our expression into the difference of two sums starting at 1:
\begin{align*} &= \left[1^3+2^3+3^3+4^3+5^3+6^3+7^3+\cdots +95^3\right]\\ &\hskip0.5in-\left[1^3+2^3+3^3+4^3\right]\\ &=\displaystyle\sum_{i=1}^{95} i^3 - \displaystyle\sum_{i=1}^4 i^3\\ &=\left[\frac{1}{2}(95)(96)\right]^2-\left[\frac{1}{2}(4)(5)\right]^2\\ &=20,793,500\,. \end{align*} -
Notice every two terms cancel with each other, since the sum is \((-1)+(+1)\text{,}\) etc. Then the terms \(n=1\) through \(n=10\) cancel, and we’re left only with the final term, \((-1)^{11}=-1\text{.}\)Written out more explicitly:\begin{align*} &\displaystyle\sum_{n=1}^{11} (-1)^{n}=-1+1-1+1-1+1-1+1-1+1-1\\ &=[-1+1]+[-1+1]+[-1+1]+[-1+1]+[-1+1]-1\\ &=0+0+0+0+0-1=-1. \end{align*}
- For every integer \(n\text{,}\) \(2n+1\) is odd, so \((-1)^{2n+1}=-1\text{.}\) Then \(\displaystyle\sum_{n=2}^{11} (-1)^{2n+1} =\displaystyle\sum_{n=2}^{11} -1 =-10\text{.}\)
1.1.8.11.
Solution.
1.1.8.12. (✳).
Solution.
- To get the limits of summation to match the given sum, we need \(n=4\text{.}\)
- Then to get the factor multiplying \(f\) to match that in the given sum, we need \(\frac{b-a}{n}=1\text{,}\) so \(b-a=4\text{.}\)
- Finally, to get the argument of \(f\) to match that in the given sum, we need\begin{gather*} a+(k-1)\frac{b-a}{n}=a-\frac{b-a}{n} +k\frac{b-a}{n}=1+k \end{gather*}Subbing in \(n=4\) and \(b-a=4\) gives \(a-1 +k=1+k\text{,}\) so \(a=2\) and \(b=6\text{.}\)
1.1.8.13.
Solution.
-
Solution 1:
- Because the index runs from \(1\) to \(3\text{,}\) there are three intervals: \(n=3\text{.}\)
- Looking at our sum, it seems reasonable to interpret \(\Delta x = 2\text{.}\) Then, since \(n=3\text{,}\) we conclude \(\frac{b-a}{3}=2\text{,}\) hence \(b-a=6\text{.}\)
- If \(\Delta x = 2\text{,}\) then \(f(x_i^*)=\left(5+2i\right)^2\text{.}\) Recall that \(x_i^*\) is the \(x\)-coordinate we use to decide the height of the \(i\)th rectangle. In a right Riemann sum, \(x_i^* = a+i\cdot\Delta x\text{.}\) So, using \(2=\Delta x\text{,}\) we can let \(f(x_i^*)=f(a+2i)=\left(5+2i\right)^2\text{.}\) This fits with the function \(f(x)=x^2\text{,}\) and \(a=5\text{.}\)
- Since \(b-a=6\text{,}\) and \(a=5\text{,}\) this tells us \(b=11\)
To sum up, we can interpret the Riemann sum as a right Riemann sum, with three intervals, of the function \(f(x)=x^2\) from \(x=5\) to \(x=11\text{.}\) - Solution 2: We could have chosen a different value for \(\Delta x\text{.}\)
- The index of the sum runs from 1 to 3, so we have \(n=3\text{.}\)
- We didn’t have to interpret \(\Delta x\) as 2--that was just the path of least resistance. We could have chosen it to be any other number--for the sake of argument, let’s say \(\Delta x=10\text{.}\) (Positive numbers are easiest to interpret, but negatives are technically allowed as well.)
- Then \(10=\frac{b-a}{n}=\frac{b-a}{3}\text{,}\) so \(b-a=30\text{.}\)
- Let’s use the paradigm of a right Riemann sum, and match up the terms of the sum given in the problem to the terms in the definition:\begin{align*} \Delta x \cdot f\left(a+i\cdot \Delta x\right)&= 2\cdot\left(5+2i\right)^2\\ 10 \cdot f(a+10i)&= 2\cdot\left(5+2i\right)^2\\ f(a+10i)&=\frac{1}{5}\cdot\left(5+2i\right)^2\\ f(a+10i)&=\frac{1}{5}\cdot\left(5+\frac{1}{5}\cdot 10i\right)^2 \end{align*}
- The easiest value of \(a\) in this case is \(a=0\text{.}\) Then \(f(\textcolor{red}{10i}) = \frac{1}{5}\cdot\left(5+\frac{1}{5}\cdot \textcolor{red}{10i}\right)^2\text{,}\) so \(f(\textcolor{red}{x})= \frac{1}{5}\cdot\left(5+\frac{1}{5}\cdot \textcolor{red}{x}\right)^2\text{.}\)
- If \(a=0\) and \(b-a=30\text{,}\) then \(b=30\text{.}\)
-
To sum up: \(n=3\text{,}\) \(a=0\text{,}\) \(b=30\text{,}\) \(\Delta x = 10\text{,}\) and \(f(x)= \frac{1}{5}\cdot\left(5+\frac{x}{5} \right)^2\text{.}\)By changing \(\Delta x\text{,}\) we changed the widths of the rectangles. The rectangles in this picture are wider and shorter than the rectangles in Solution 1. Their areas are the same: 98, 162, and 242.
- Solution 3: We could have chosen a different value of \(a\text{.}\)
- Suppose \(\Delta x = 2\text{,}\) and we interpret our sum as a right Riemann sum, but we didn’t assume \(a=5\text{.}\) We could have chosen \(a\) to be any number--say, \(a=1\text{.}\)
- Let’s match up what we’re given in the problem to what we’re given as a definition:\begin{align*} \Delta x \cdot f\left(a+i\cdot\Delta x\right)&=2\cdot\left(5+2i\right)^2\\ 2 \cdot f\left(1+2i\right)&=2\cdot\left(5+2i\right)^2\\ f\left(1+2i\right)&=\left(5+2i\right)^2\\ f\left(1+2i\right)&=\left(4+1+2i\right)^2 \end{align*}
- Since \(f(\textcolor{red}{1+2i})=\left(4+\textcolor{red}{1+2i}\right)^2\text{,}\) we have \(f(\textcolor{red}{x})=\left(4+\textcolor{red}{x}\right)^2\)
- Since \(a=1\) and \(\frac{b-a}{3}=2\text{,}\) in this case \(b=7\text{.}\)
-
To sum up: \(n=3\text{,}\) \(a=1\text{,}\) \(b=7\text{,}\) \(\Delta x=2\text{,}\) and \(f(x)=(4+x)^2\text{.}\)This picture is a lot like the picture in Solution 1, but shifted to the left. By changing \(a\text{,}\) we changed the left endpoint of our region.
- Solution 4: We could have chosen a different kind of Riemann sum.
- We didn’t have to assume that we were dealing with a right Riemann sum. Suppose \(\Delta x =2\text{,}\) and we have a midpoint Riemann sum.
- Let’s match up what we’re given in the problem with what we’re given in the definition:\begin{align*} \Delta x \cdot f\left(a+\left(i-\tfrac{1}{2}\right)\Delta x\right)&=2\cdot\left(5+2i\right)^2\\ 2 \cdot f\left(a+\left(i-\tfrac{1}{2}\right)2\right)&=2\cdot\left(5+2i\right)^2\\ f\left(a+\left(i-\tfrac{1}{2}\right)2\right)&=\left(5+2i\right)^2\\ f\left(a+2i-1\right)&=\left(5+2i\right)^2\\ f\left((a-1)+2i\right)&=\left(5+2i\right)^2 \end{align*}
- It is now convenient to set \(a-1=5\text{,}\) hence \(a=6\text{.}\)
- Then \(f(\textcolor{red}{5+2i})=(\textcolor{red}{5+2i})^2\text{,}\) so \(f(\textcolor{red}{x})=\textcolor{red}{x}^2\)
- Since \(2=\frac{b-a}{3}\) and \(a=6\text{,}\) we see \(b=12\text{.}\)
-
To sum up: \(n=3\text{,}\) \(a=6\text{,}\) \(b=12\text{,}\) \(\Delta x = 2\text{,}\) and \(f(x)=x^2\text{.}\)By choosing to interpret our sum as a midpoint Riemann sum instead of a right Riemann sum, we changed where our rectangles intersect the graph \(y=f(x)\text{:}\) instead of the graph hitting the right corner of the rectangle, it hits in the middle.
1.1.8.14.
Solution.
- Since the sum has five terms (\(i\) runs from 1 to 5), there are 5 rectangles. That is, \(n=5\text{.}\)
- In the definition of the Riemann sum, note that the term \(\Delta x\) appears twice: once multiplied by the entire term, and once multiplied by \(i-1\text{.}\) So, a convenient choice for \(\Delta x\) is \(\frac{\pi}{20}\text{,}\) because this is the constant that is both multiplied at the start of the term, and multiplied by \(i-1\text{.}\)
- Since \(\dfrac{\pi}{20}=\Delta x = \dfrac{b-a}{n} = \dfrac{b-a}{5}\text{,}\) we see \(b-a=\dfrac{5\pi}{20}=\dfrac{\pi}{4}\text{.}\)
- We match the terms in the definition with the terms in the problem:\begin{align*} f(a+(i-1)\Delta x) & = \tan\left(\frac{\pi (i-1)}{20}\right)\\ f\left(a+(i-1)\frac{\pi}{20}\right) & = \tan\left((i-1)\frac{\pi }{20}\right) \end{align*}So, we choose \(a=0\) and \(f(x) = \tan x\text{.}\)
- Since \(a=0\) and \(b-a=\frac{\pi}{4}\text{,}\) we see \(b=\frac{\pi}{4}\text{.}\)
1.1.8.15. (✳).
Solution.
-
Option 1: right Riemann sum. If our sum is a right Riemann sum, then we take the heights of the rectangles from the right endpoint of each interval.Then \(a=0.5\) and \(b=4.5\text{.}\) Therefore: \(\sum\limits_{k=0}^3 f (1.5 + k) \cdot 1\) is a right Riemann sum on the interval \([0.5,4.5]\) with \(n=4\text{.}\)
-
Option 2: left Riemann sum. If our sum is a left Riemann sum, then we take the heights of the rectangles from the left endpoint of each interval.Then \(a=1.5\) and \(b=5.5\text{.}\) Therefore: \(\sum\limits_{k=0}^3 f (1.5 + k) \cdot 1\) is a left Riemann sum on the interval \([1.5,5.5]\) with \(n=4\text{.}\)
-
Option 3: midpoint Riemann sum. If our sum is a midpoint Riemann sum, then we take the heights of the rectangles from the midpoint of each interval.Then \(a=1\) and \(b=5\text{.}\) Therefore: \(\sum\limits_{k=0}^3 f (1.5 + k) \cdot 1\) is a midpoint Riemann sum on the interval \([1,5]\) with \(n=4\text{.}\)
1.1.8.16.
Solution.
1.1.8.17.
Solution.
1.1.8.18. (✳).
Solution.
1.1.8.19. (✳).
Solution.
1.1.8.20. (✳).
Solution.
1.1.8.21. (✳).
Solution.
1.1.8.22. (✳).
Solution.
1.1.8.23. (✳).
Solution.
1.1.8.24. (✳).
Solution.
1.1.8.25. (✳).
Solution.
- Choice #1: If we set \(\De x = \frac{2}{n}\) and \(x_i^*= \frac{2i}{n}\text{,}\) i.e. \(x_i^* = a + i\De x\) with \(a=0\text{,}\) then\begin{align*} \lim_{n\rightarrow\infty} \bigg( \sum_{i=1}^n e^{-1-2i/n}\cdot \frac{2}{n} \bigg) &=\lim_{n\rightarrow\infty} \bigg( \sum_{i=1}^n e^{-1-x_i^*}\De x \bigg)\\ &=\lim_{n\rightarrow\infty} \bigg( \sum_{i=1}^n f(x_i^*)\De x \bigg)\\ &\hskip0.25in\text{with $f(x) = e^{-1-x}$}\\ &=\int_a^b f(x)\,\,\dee{x}\\ &\hskip0.25in\text{with $a=x_0=0$ and $b=x_n=2$}\\ &=\int_0^2 e^{-1-x}\ \,\dee{x} \end{align*}
- Choice #2: If we set \(\De x = \frac{2}{n}\) and \(x_i^*= 1+\frac{2i}{n}\text{,}\) i.e. \(x_i^* = a + i\De x\) with \(a=1\text{,}\) then\begin{align*} \lim_{n\rightarrow\infty} \bigg( \sum_{i=1}^n e^{-1-2i/n}\cdot \frac{2}{n} \bigg) &=\lim_{n\rightarrow\infty} \bigg( \sum_{i=1}^n e^{-x_i^*}\De x \bigg) \\ &=\lim_{n\rightarrow\infty} \bigg( \sum_{i=1}^n f(x_i^*)\De x \bigg)\\ &\hskip0.25in\text{with $f(x) = e^{-x}$}\\ &=\int_a^b f(x)\,\,\dee{x}\\ &\hskip0.25in\text{with $a=x_0=1$ and $b=x_n=3$}\\ &=\int_1^3 e^{-x}\ \,\dee{x} \end{align*}
- Choice #3: If we set \(\De x = \frac{1}{n}\) and \(x_i^*= \frac{i}{n}\text{,}\) i.e. \(x_i^* = a + i\De x\) with \(a=0\text{,}\) then\begin{align*} \lim_{n\rightarrow\infty} \bigg( \sum_{i=1}^n e^{-1-2i/n}\cdot \frac{2}{n} \bigg) &=\lim_{n\rightarrow\infty} \bigg( \sum_{i=1}^n e^{-1-2x_i^*}\ 2\De x \bigg) \\ &=\lim_{n\rightarrow\infty} \bigg( \sum_{i=1}^n f(x_i^*)\De x \bigg)\\ &\hskip0.25in\text{with $f(x) = 2e^{-1-2x}$}\\ &=\int_a^b f(x)\,\,\dee{x}\\ &\hskip0.25in\text{with $a=x_0=0$ and $b=x_n=1$}\\ &=2\int_0^1 e^{-1-2x}\ \,\dee{x} \end{align*}
- Choice #4: If we set \(\De x = \frac{1}{n}\) and \(x_i^*= \frac{1}{2}+\frac{i}{n}\text{,}\) i.e. \(x_i = a + i\De x\) with \(a=\frac{1}{2}\text{,}\) then\begin{align*} \lim_{n\rightarrow\infty} \bigg( \sum_{i=1}^n e^{-1-2i/n}\cdot \frac{2}{n} \bigg) &=\lim_{n\rightarrow\infty} \bigg( \sum_{i=1}^n e^{-2x_i}\ 2\De x \bigg) \\ &=\lim_{n\rightarrow\infty} \bigg( \sum_{i=1}^n f(x_i^*)\De x \bigg)\\ &\hskip0.25in\text{with $f(x) = 2e^{-2x}$}\\ &=\int_a^b f(x)\,\,\dee{x}\\ &\hskip0.25in\text{with $a=x_0=\frac{1}{2}$ and $b=x_n=\frac{3}{2}$}\\ &=2\int_{1/2}^{3/2} e^{-2x}\ \,\dee{x} \end{align*}
1.1.8.26.
Solution.
1.1.8.27.
Solution.
- Solution 1: If we factor out \(r^5\text{,}\) then what’s left fits the form of Equation 1.1.3:\begin{align*} r^5+r^6+r^7+\cdots+r^{100}&=r^5\left[1+r+r^2+\cdots + r^{95}\right]\\ &= r^5\left(\frac{r^{96}-1}{r-1}\right)\ . \end{align*}
- Solution 2: We know how to evaluate sums of this form if they start at 1, so we re-write our sum as follows:\begin{align*} r^5+r^6+r^7+\cdots+r^{100}&=\left(1+r+r^2+r^3+r^4+r^5+\cdots+r^{100}\right) \\ &\hskip0.25in-\left(1+r+r^2+r^3+r^4\right)\\ &=\frac{r^{101}-1}{r-1} - \frac{r^5-1}{r-1}\\ &=\frac{r^{101}-1-r^5+1}{r-1}=\frac{r^{101}-r^5}{r-1}\\ &=r^5\left(\frac{r^{96}-1}{r-1}\right)\ . \end{align*}
1.1.8.28. (✳).
Solution.
- at the left hand end of the domain of integration \(x=-1\) and the integrand \(|2x|=|-2|=2\) and
- as \(x\) increases from \(-1\) towards \(0\text{,}\) the integrand \(|2x|=-2x\) decreases linearly, until
- when \(x\) hits \(0\) the integrand hits \(|2x|=|0|=0\) and then
- as \(x\) increases from \(0\text{,}\) the integrand \(|2x|=2x\) increases linearly, until
- when \(x\) hits \(+2\text{,}\) the right hand end of the domain of integration, the integrand hits \(|2x|=|4|=4\text{.}\)
1.1.8.29.
Solution.
1.1.8.30.
Solution.
1.1.8.31.
Solution.
1.1.8.32.
Solution.
1.1.8.33. (✳).
Solution.
- \(\int_0^1 f(x)\,\,\dee{x}\text{:}\) a right-angled triangle of height \(1\) and base \(1\) and hence area \(0.5\text{.}\)
- \(\int_1^3 f(x)\,\,\dee{x}\text{:}\) a rectangle of height \(1\) and base \(2\) and hence area \(2\text{.}\)
1.1.8.34. (✳).
Solution.
1.1.8.35.
Solution.
1.1.8.36.
Solution.
1.1.8.37. (✳).
Solution.
-
Solution #1: Set \(x_i^*=-2+\frac{2i}{n}\text{.}\) Then \(a=x_0=-2\) and \(b=x_n=0\) and \(\Delta x=\frac{2}{n}\text{.}\) So\begin{align*} &\lim_{n\rightarrow\infty} \sum_{i=1}^n\frac{2}{n}\sqrt{4-\left(-2+\frac{2i}{n}\right)^2}\\ &= \lim_{n\rightarrow\infty} \sum_{i=1}^n f(x_i^*)\Delta x\qquad \text{ with $f(x) = \sqrt{4-x^2}$ and $\Delta x = \frac{2}{n}$ }\\ &=\int_{-2}^0 \sqrt{4-x^2}\,\,\dee{x} \end{align*}For the integral \(\int_{-2}^0 \sqrt{4-x^2}\,\,\dee{x}\text{,}\) \(y=\sqrt{4-x^2}\) is equivalent to \(x^2+y^2=4\text{,}\) \(y\ge 0\text{.}\) So the integral represents the area between the upper half of the circle \(x^2+y^2=4\) (which has radius \(2\)) and the \(x\)-axis with \(-2\le x\le 0\text{,}\) which is a quarter circle with area \(\frac{1}{4}\cdot \pi\, 2^2 = \pi\text{.}\)
-
Solution #2: Set \(x_i^*=\frac{2i}{n}\text{.}\) Then \(a=x_0=0\) and \(b=x_n=2\) and \(\Delta x=\frac{2}{n}\text{.}\) So\begin{align*} &\lim_{n\rightarrow\infty} \sum_{i=1}^n\frac{2}{n}\sqrt{4-\left(-2+\frac{2i}{n}\right)^2}\\ &= \lim_{n\rightarrow\infty} \sum_{i=1}^n f(x_i^*)\Delta x\quad \text{ with $f(x) = \sqrt{4-(-2+x)^2}$, $\Delta x = \frac{2}{n}$ }\\ &=\int_0^2 \sqrt{4-(-2+x)^2}\,\,\dee{x} \end{align*}For the integral \(\int_{0}^2 \sqrt{4-(-2+x)^2}\,\,\dee{x}\ , y=\sqrt{4-(x-2)^2}\) is equivalent to \((x-2)^2+y^2=4\text{,}\) \(y\ge 0\text{.}\) So the integral represents the area between the upper half of the circle \((x-2)^2+y^2=4\) (which is centered at \((2,0)\) and has radius \(2\)) and the \(x\)-axis with \(0\le x\le 2\text{,}\) which is a quarter circle with area \(\frac{1}{4}\cdot \pi\, 2^2 = \pi\text{.}\)
1.1.8.38. (✳).
Solution.
1.1.8.39. (✳).
Solution.
1.1.8.40. (✳).
Solution.
1.1.8.41. (✳).
Solution.
1.1.8.42.
Solution.
Since \(\Delta x = 10\) and \(a=-5\text{,}\)
\begin{align*} \displaystyle\sum_{i=1}^{10} 3(7+2i)^2\sin(4i) &= \displaystyle\sum_{i=1}^{10} 10 f(-5+10i)\\ \end{align*}Dividing both expressions by 10,
\begin{align*} \displaystyle\sum_{i=1}^{10} \frac{3}{10}(7+2i)^2\sin(4i) &= \displaystyle\sum_{i=1}^{10} f(-5+10i)\\ \end{align*}So, we have an expression for \(f(-5+10i)\text{:}\)
\begin{align*} f(-5+10i) &= \frac{3}{10}(7+2i)^2\sin(4i)\\ \end{align*}In order to find \(f(x)\text{,}\) let \(x=-5+10i\text{.}\) Then \(i=\frac{x}{10}+\frac{1}{2}\text{.}\)
\begin{align*} f(x) &= \frac{3}{10}\left(7+2\left(\frac{x}{10}+\frac{1}{2}\right)\right)^2\sin\left(4\left(\frac{x}{10}+\frac{1}{2}\right)\right)\\ &=\frac{3}{10}\left(\frac{x}{5}+8\right)^2\sin\left(\frac{2x}{5}+2\right)\ . \end{align*}1.1.8.43.
Solution.
The sum in parenthesis has the form of a geometric sum, with \(r=2^{1/n}\text{:}\)
\begin{align*} &=\lim_{n \to \infty}\frac{2^{1/n}}{n}\left( \frac{\left(2^{1/n}\right)^n-1}{2^{1/n}-1} \right)\\ &=\lim_{n \to \infty}\frac{2^{1/n}}{n}\left( \frac{2-1}{2^{1/n}-1} \right)\\ &=\lim_{n \to \infty} \frac{2^{1/n}}{n(2^{1/n}-1)}\\ \end{align*}Note as \(n \to \infty\text{,}\) \(1/n \to 0\text{,}\) so the numerator has limit 1, while the denominator has indeterminate form \(\infty\cdot 0\text{.}\) So, we’ll do a little algebra to get this into a l’Hôpital-style indeterminate form:
\begin{align*} &=\lim_{n \to \infty} \frac{\frac{1}{n}\cdot2^{1/n}}{2^{1/n}-1}\\ &=\lim_{n \to \infty} \underbrace{\frac{\frac{1}{n}}{1-2^{-1/n}}}_{\atp{\mathrm{num}\to 0}{\mathrm{den}\to 0}}\\ \end{align*}Now we can use l’Hôpital’s rule. Recall \(\diff{}{x}\left\{2^x\right\}=2^x\log x\text{,}\) where \(\log x\) is the natural logarithm of \(x\text{,}\) also sometimes written \(\ln x\text{.}\) We’ll need to use the chain rule when we differentiate the denominator.
\begin{align*} &=\lim_{n \to \infty} \frac{\frac{-1}{n^2}}{-2^{-1/n}\log 2 \cdot \frac{1}{n^2}}\\ &=\lim_{n \to \infty} \frac{2^{1/n}}{\log 2}\\ &=\frac{1}{\log 2} \end{align*}1.1.8.44.
Solution.
Now the sum in parentheses has the form of a geometric sum, with \(r=10^{\frac{b-a}{n}}\text{:}\)
\begin{align*} &=\lim_{n \to \infty} \frac{b-a}{n}\cdot 10^{a}\cdot 10^{\frac{b-a}{n}}\left( \frac{\left(10^{\frac{b-a}{n}}\right)^n-1}{10^{\frac{b-a}{n}}-1} \right)\\ &=\lim_{n \to \infty} \frac{\textcolor{blue}{b-a}}{n}\cdot \textcolor{red}{10^{a}}\cdot 10^{\frac{b-a}{n}}\left( \frac{\textcolor{purple}{10^{b-a}-1}}{10^{\frac{b-a}{n}}-1} \right)\\ \end{align*}The coloured parts do not depend on \(n\text{,}\) so for simplicity we can move them outside the limit.
\begin{align*} &=\textcolor{blue}{(b-a)}\cdot\textcolor{red}{10^a}\left(\textcolor{purple}{10^{b-a}-1}\right)\lim_{n \to \infty} \frac{1}{n}\cdot \left( \frac{ 10^{\frac{b-a}{n}} }{10^{\frac{b-a}{n}}-1} \right)\\ &={(b-a)}\cdot\left({10^{b}-10^a}\right)\lim_{n \to \infty} \underbrace{\left( \frac{ 1/n}{1-10^{-\frac{b-a}{n}}} \right)}_{\atp{\mathrm{num}\to 0}{\mathrm{den}\to 0}}\\ \end{align*}Now we can use l’Hôpital’s rule. Recall \(\diff{}{x}\left\{10^x\right\}=10^x\log x\text{,}\) where \(\log x\) is the natural logarithm of \(x\text{,}\) also sometimes written \(\ln x\text{.}\) For the denominator, we will have to use the chain rule.
\begin{align*} &={(b-a)}\cdot\left({10^{b}-10^a}\right)\lim_{n \to \infty} \left( \frac{ -1/n^2}{-10^{-\frac{b-a}{n}}\cdot \log 10 \cdot \frac{b-a}{n^2}} \right)\\ &={(b-a)}\cdot\left({10^{b}-10^a}\right)\lim_{n \to \infty} \left( \frac{ 1}{10^{-\frac{b-a}{n}}\cdot \log 10 \cdot (b-a)} \right)\\ &={(b-a)}\cdot\left({10^{b}-10^a}\right) \left( \frac{ 1}{ \log 10 \cdot (b-a)} \right)\\ &=\frac{1}{\log 10}\left(10^b-10^a\right) \end{align*}1.1.8.45.
Solution.
-
Area of sector: The sector is a portion of a circle with radius 1, with inner angle \(\theta\text{.}\) So, its area is \(\frac{\theta}{2\pi}\left(\mbox{area of circle}\right) = \frac{\theta}{2\pi}\left(\pi\right) = \frac{\theta}{2}\text{.}\)Our job now is to find \(\theta\) in terms of \(a\text{.}\) Note \(\frac{\pi}{2}-\theta\) is the inner angle of the red triangle, which lies in the unit circle. So, \(\cos\left(\frac{\pi}{2}-\theta\right)=a\text{.}\) Then \(\frac{\pi}{2}-\theta= \arccos(a)\text{,}\) and so \(\theta = \frac{\pi}{2} - \arccos(a)\text{.}\)Then the area of the sector is \(\frac{\pi}{4} - \frac{1}{2}\arccos(a)\) square units.
- Area of triangle: The triangle has base \(a\text{.}\) Its height is the \(y\)-value of the function when \(x=a\text{,}\) so its height is \(\sqrt{1-a^2}\text{.}\) Then the area of the triangle is \(\frac{1}{2}a\sqrt{1-a^2}\text{.}\)
1.1.8.46.
Solution.
-
The difference between our upper and lower bounds is the difference in areas between the larger set of rectangles and the smaller set of rectangles. Drawing them on a single picture makes this a little clearer.Each of the rectangles has width \(\frac{b-a}{n}\text{,}\) since we took a segment of the \(x\)-axis with length \(b-a\) and chopped it into \(n\) pieces. We could calculate the height of each rectangle, but it would be a little complicated, since it differs for each of them. An easier method is to notice that the area we want to calculate can be imagined as a single rectangle:The rectangle has base \(\frac{b-a}{n}\text{.}\) Its highest coordinate is \(f(a)\text{,}\) and its lowest is \(f(b)\text{,}\) so its height is \(f(b)-f(a)\text{.}\) Therefore, the difference in area between our lower bound and our upper bound is:\begin{equation*} \left[f(b)-f(a)\right]\cdot\frac{b-a}{n} \end{equation*}
-
We want to give a range with length at most 0.01, and guarantee that the area under the curve \(y=f(x)\) is inside that range. In the previous part, we figured out that when we use \(n\) rectangles, the length of our range is \(\left[f(b)-f(a)\right]\cdot\frac{b-a}{n}\text{.}\) So, all we have to do is set this to be less than or equal to 0.01, and solve for \(n\text{:}\)\begin{align*} \left[f(b)-f(a)\right]\cdot\frac{b-a}{n}&\leq 0.01\\ 100\left[f(b)-f(a)\right]\cdot(b-a)&\leq n \end{align*}We can choose \(n\) to be an integer that is greater than or equal to \(100\left[f(b)-f(a)\right]\cdot(b-a)\text{.}\) Using that many rectangles, we find an upper and lower bound for the area under the curve. If we choose any number between our upper and lower bound as an approximation for the area under the curve, our error is no more than 0.01.
1.1.8.47.
Solution.
- If we are using a left Riemann sum, our rectangle has height \(f(x)=mx+c\text{.}\) Then it has area \(w(mx+c)\text{.}\)
- If we are using a right Riemann sum, our rectangle has height \(f(x+w)=m(x+w)+c=mx+c+mw\text{.}\) Then it has area \(w(mx+c+mw)\text{.}\)
- If we are using a midpoint Riemann sum, our rectangle has height \(f(x+\frac{1}{2}w)=m(x+\frac{1}{2}w)+c=mx+c+\frac{1}{2}mw\text{.}\) Then it has area \(w\left(mx+c+\frac{1}{2}w\right)\text{.}\)
1.2 Basic properties of the definite integral
1.2.3 Exercises
1.2.3.1.
Solution.
-
\(\displaystyle\int_a^a f(x)\,\dee{x}=0\)The area under the curve is zero, because it’s a region with no width.
-
\(\displaystyle\int_a^b f(x)\,\dee{x}=\textcolor{blue}{ \displaystyle\int_a^c f(x)\,\dee{x}} +\textcolor{red}{ \int_c^b f(x)\dee{x} }\)If we assume \(a \leq c \leq b\text{,}\) then this identity simply tells us that if we add up the area under the curve from \(a\) to \(c\text{,}\) and from \(c\) to \(b\text{,}\) then we get the whole area under the curve from \(a\) to \(b\text{.}\)(The situation is slightly more complicated when \(c\) is not between \(a\) and \(b\text{,}\) but it still works out.)
-
\(\displaystyle\int_a^b \left( f(x) + g(x) \right)\,\dee{x} =\textcolor{blue}{ \displaystyle\int_a^b f(x)\,\dee{x}} +\textcolor{red}{ \displaystyle\int_a^b g(x)\,\dee{x}}\)The blue-shaded area in the picture above is \(\displaystyle\int_a^b f(x)\dee{x}\text{.}\) The area under the curve \(f(x)+g(x)\) but above the curve \(f(x)\) (shown in red) is \(\displaystyle\int_a^b g(x)\dee{x}\text{.}\)
1.2.3.2.
Solution.
we see
\begin{align*} \int\limits_a^b\cos x\dee{x} &= \int\limits_a^0 \cos x\dee{x}+ \int\limits_0^b \cos x\dee{x}\\ &= -\int\limits_0^a \cos x\dee{x}+ \int\limits_0^b \cos x\dee{x}\\ &=-\sin a + \sin b\\ &=\sin b - \sin a \end{align*}1.2.3.3. (✳).
Solution.
1.2.3.4.
Solution.
-
\(\Delta x = \dfrac{b-a}{n}=\dfrac{0-5}{100} = -\dfrac{1}{20}\)Note: if we were to use the Riemann-sum definition of a definite integral, this is how we would justify the identity \(\int\limits_a^b f(x)\dee{x}=-\int\limits_b^a f(x)\dee{x}\text{.}\)
- The heights of the rectangles are given by \(f(x_i)\text{,}\) where \(x_i = a+i\Delta x = 5 - \frac{i}{20}\text{.}\) Since \(f(x)\) only gives positive values, \(f(x_i) \gt 0\text{,}\) so the heights of the rectangles are positive.
- Our Riemann sum is the sum of the signed areas of individual rectangles. Each rectangle has a negative base (\(\Delta x\)) and a positive height (\(f(x_i)\)). So, each term of our sum is negative. If we add up negative numbers, the sum is negative. So, the Riemann sum is negative.
- Since \(f(x)\) is always above the \(x\)-axis, \(\int\limits_0^5 f(x)\dee{x}\) is positive.
1.2.3.5. (✳).
Solution.
1.2.3.6. (✳).
Solution.
1.2.3.7. (✳).
Solution.
1.2.3.8.
Solution.
- Since \(\sqrt{1-x^2}\) is an even function,\begin{align*} \int_{a}^0 \sqrt{1-x^2}\dee{x} &=\int_{0}^{|a|} \sqrt{1-x^2}\dee{x} \\ &= \frac{\pi}{4} - \frac{1}{2}\arccos(|a|)+\frac{1}{2}|a|\sqrt{1-|a|^2}\\ &=\frac{\pi}{4} - \frac{1}{2}\arccos(-a)-\frac{1}{2}a\sqrt{1-a^2} \end{align*}Alternatively, since \(\arccos(-a) = \pi-\arccos(a)\) we also have\begin{align*} \int_{a}^0 \sqrt{1-x^2}\ \dee{x} \amp=-\frac{\pi}{4} + \frac{1}{2}\arccos(a)-\frac{1}{2}a\sqrt{1-a^2} \end{align*}
- Note \(\displaystyle\int_{0}^1 \sqrt{1-x^2}\dee{x}=\frac{\pi}{4}\text{,}\) since the area under the curve represents one-quarter of the unit circle. Then,\begin{align*} \displaystyle\int_{a}^1 \sqrt{1-x^2}\dee{x}&= \displaystyle\int_{0}^1 \sqrt{1-x^2}\dee{x}- \displaystyle\int_{0}^a \sqrt{1-x^2}\dee{x}\\ &=\frac{\pi}{4}-\left(\frac{\pi}{4} - \frac{1}{2}\arccos(a)+\frac{1}{2}a\sqrt{1-a^2}\right)\\ &=\frac{1}{2}\arccos(a)-\frac{1}{2}a\sqrt{1-a^2} \end{align*}
1.2.3.9. (✳).
Solution.
1.2.3.10.
Solution.
1.2.3.11.
Solution.
Also,
\begin{align*} \int_{-2}^2 f(x)\dee{x}&=\int_{-2}^0 f(x)\dee{x}+\int_{0}^2 f(x)\dee{x}\\ \end{align*}So,
\begin{align*} \int_{-2}^0 f(x)\dee{x}&=\int_{-2}^2 f(x)\dee{x}-\int_{0}^2 f(x)\dee{x}\\ &=10-5=5 \end{align*}1.2.3.12. (✳).
Solution.
1.2.3.13. (✳).
Solution.
1.2.3.14. (✳).
Solution.
1.2.3.15.
Solution.
1.2.3.16.
Solution.
- \begin{align*} (ax)^2+(by)^2&=1\\ by&=\sqrt{1-(ax)^2}\\ y &= \frac{1}{b}\sqrt{1-(ax)^2} \end{align*}
- The values of \(x\) in the domain of the function above are those that satisfy \(1-(ax)^2 \geq 0\text{.}\) That is, \(-\frac{1}{a}\leq x \leq \frac{1}{a}\text{.}\) Therefore, the upper half of the ellipse has area\begin{align*} \displaystyle\frac{1}{b}&\displaystyle\int_{-\frac{1}{a}}^{\frac{1}{a}}\sqrt{1-(ax)^2}\dee{x}\\ \end{align*}
The upper half of a circle has equation \(y=\sqrt{r^2-x^2}\text{.}\)
\begin{align*} &=\frac{1}{b}\int_{-\frac{1}{a}}^{\frac{1}{a}}\sqrt{a^2\left(\frac{1}{a^2}-x^2\right)}\dee{x}\\ &=\frac{1}{b}\int_{-\frac{1}{a}}^{\frac{1}{a}}a\sqrt{\frac{1}{a^2}-x^2}\dee{x}\\ &=\frac{a}{b}\int_{-\frac{1}{a}}^{\frac{1}{a}}\sqrt{\frac{1}{a^2}-x^2}\dee{x} \end{align*} -
The function \(y=\sqrt{\dfrac{1}{a^2}-x^2}\) is the upper-half of the circle centred at the origin with radius \(\dfrac{1}{a}\text{.}\) So, the expression from (b) evaluates to \(\left(\dfrac{a}{b}\right)\dfrac{\pi}{2a^2} = \dfrac{\pi}{2ab}\text{.}\)The expression from (b) was half of the ellipse, so the area of the ellipse is \(\dfrac{\pi}{ab}\text{.}\)
- The area of the unit circle \(x^2+y^2=1\) is \(\pi \text{.}\)
- The ellipse \((ax)^2+y^2=1\) is obtained by shrinking the unit circle horizontally by a factor of \(a\text{.}\) So, its area is \(\dfrac{\pi}{a}\text{.}\)
- Further, the ellipse \((ax)^2+(by)^2=1\) is obtained from the previous ellipse by shrinking it vertically by a factor of \(b\text{.}\) So, its area is \(\dfrac{\pi}{ab}\text{.}\)
1.2.3.17.
Solution.
- even \(\times\) even: If \(f\) and \(g\) are both even, then \(h(-x)=f(-x)\cdot g(-x) = f(x)\cdot g(x)=h(x)\text{,}\) so their product is even.
- odd \(\times\) odd: If \(f\) and \(g\) are both odd, then \(h(-x)=f(-x)\cdot g(-x) =[- f(x)]\cdot [-g(x)]=f(x)\cdot g(x)=h(x)\text{,}\) so their product is even.
- even \(\times\) odd: If \(f\) is even and \(g\) is odd, then \(h(-x)=f(-x)\cdot g(-x) = f(x)\cdot[- g(x)]=-[f(x)\cdot g(x)]=-h(x)\text{,}\) so their product is odd. Because multiplication is commutative, the order we multiply the functions in doesn’t matter.
1.2.3.18.
Solution.
1.2.3.19.
Solution.
- \(f(x)=f(-x)\) (since \(f(x)\) is even), and
- \(f(x)=-f(-x)\) (since \(f(x)\) is odd).
- So, \(f(x)=-f(x)\text{.}\)
- Then (adding \(f(x)\) to both sides) we see \(2f(x)=0\text{,}\) so \(f(x)=0\text{.}\)
1.2.3.20.
Solution.
-
Solution 1: Suppose \(f(x)\) is an odd function. We investigate \(f'(x)\) using the chain rule:\begin{align*} f(-x)&=-f(x)& \mbox{(odd function)}\\ \diff{}{x}\{f(-x)\}&=\diff{}{x}\{-f(x)\}\\ -f'(-x)&=-f'(x) & \mbox{(chain rule)}\\ f'(-x)&=f'(x) \end{align*}So, when \(f(x)\) is odd, \(f'(x)\) is even.Similarly, suppose \(f(x)\) is even.\begin{align*} f(-x)&=f(x)& \mbox{(even function)}\\ \diff{}{x}\{f(-x)\}&=\diff{}{x}\{f(x)\}\\ -f'(-x)&=f'(x) & \mbox{(chain rule)}\\ f'(-x)&=-f'(x) \end{align*}So, when \(f(x)\) is even, \(f'(x)\) is odd.
-
Solution 2: Another way to think about this problem is to notice that “mirroring” a function changes the sign of its derivative. Then since an even function is “mirrored once” (across the \(y\)-axis), it should have \(f'(x)=-f'(-x)\text{,}\) and so the derivative of an even function should be an odd function. Since an odd function is “mirrored twice” (across the \(y\)-axis and across the \(x\)-axis), it should have \(f'(x)=-(-f'(-x))=f'(-x)\text{.}\) So the derivative of an odd function should be even. These ideas are presented in more detail below.First, we consider the case where \(f(x)\) is even, and investigate \(f'(x)\text{.}\)The whole function has a mirror-like symmetry across the \(y\)-axis. So, at \(x\) and \(-x\text{,}\) the function will have the same “steepness,” but if one is increasing then the other is decreasing. That is, \(f'(-x)=-f'(x)\text{.}\) (In the picture above, compare the slope at some point \(a_i\) with its corresponding point \(-a_i\text{.}\)) So, \(f'(x)\) is odd when \(f(x)\) is even.Second, let’s consider the case where \(f(x)\) is odd, and investigate \(f'(x)\text{.}\) Suppose the blue graph below is \(y=f(x)\text{.}\) If \(f(x)\) were even, then to the left of the \(y\)-axis, it would look like the orange graph, which we’ll call \(y=g(x)\text{.}\)From our work above, we know that, for every \(x \gt 0\text{,}\) \(-f'(x)=g'(-x)\text{.}\) When \(x \lt 0\text{,}\) \(f(x)=-g(x)\text{.}\) So, if \(x \gt 0\text{,}\) then \(-f'(x)=g'(-x)=-f'(-x)\text{.}\) In other words, \(f'(x)=f'(-x)\text{.}\) Similarly, if \(x \lt 0\text{,}\) then \(f'(x)=-g'(x)=f'(-x)\text{.}\) Therefore \(f'(x)\) is even. (In the graph below, you can anecdotally verify that \(f'(a_i)=f'(-a_i)\text{.}\))
1.3 The Fundamental Theorem of Calculus
1.3.2 Exercises
1.3.2.1. (✳).
Solution.
1.3.2.2. (✳).
Solution.
- One function with derivative \(x^3\) is \(\dfrac{x^4}{4}\text{.}\)
- To find an antiderivative of \(\sin(2x)\text{,}\) we might first guess \(\cos(2x)\text{;}\) checking, we see \(\diff{}{x}\{\cos(2x)\}=-2\sin(2x)\text{.}\) So, we only need to multiply by \(-\dfrac{1}{2}\text{:}\) \(\displaystyle\diff{}{x}\left\{-\dfrac{1}{2}\cos 2x\right\}=\sin(2x)\text{.}\)
8
1.3.2.3. (✳).
Solution.
1.3.2.4.
Solution.
so perhaps similarly
\begin{align*} \int \frac{1}{x^2}\dee{x}&\stackrel{?}{=}\log|x^2|+C=\log(x^2)+C\\ \end{align*}We check by differentiating:
\begin{align*} \diff{}{x}\{\log(x^2)\} &= \diff{}{x}\{2\log x\}=\frac{2}{x} \neq \frac{1}{x^2} \end{align*}1.3.2.5.
Solution.
so perhaps
\begin{align*} \diff{}{x}\left\{\frac{\sin(e^x)}{e^x}\right\} &\stackrel{?}{=} \cos(e^x)\\ \end{align*}We check by differentiating:
\begin{align*} \diff{}{x}\left\{\frac{\sin(e^x)}{e^x}\right\} &=\frac{e^x\left(\cos(e^x)\cdot e^x\right)-\sin(e^x)e^x}{e^{2x}} &\mbox{(quotient rule)}\\ & = \cos (e^x) - \frac{\sin(e^x)}{e^x}\\ &\neq \cos(e^x) \end{align*}1.3.2.6.
Solution.
1.3.2.7.
Solution.
1.3.2.8.
Solution.
1.3.2.9.
Solution.
- We differentiate with respect to \(a\text{.}\) Recall \(\diff{}{x}\{\arccos x\} = \frac{-1}{\sqrt{1-x^2}}\text{.}\) To differentiate \(\frac{1}{2}a\sqrt{1-a^2}\text{,}\) we use the product and chain rules.\begin{align*} &\diff{}{a}\left\{\frac{\pi}{4} - \frac{1}{2}\arccos(a)+\frac{1}{2}a\sqrt{1-a^2}\right\}\\ &=0-\frac{1}{2}\cdot\frac{-1}{\sqrt{1-a^2}} + \left(\frac{1}{2}a\right)\cdot\frac{-2a}{2\sqrt{1-a^2}} + \frac{1}{2}\sqrt{1-a^2}\\ &=\frac{1}{2\sqrt{1-a^2}}- \frac{a^2}{2\sqrt{1-a^2}}+\frac{1-a^2}{2\sqrt{1-a^2}}\\ &=\frac{1-a^2+1-a^2}{2\sqrt{1-a^2}}\\ &=\frac{2(1-a^2)}{2\sqrt{1-a^2}}\\ &=\sqrt{1-a^2} \end{align*}
-
Let \(G(x) = \frac{\pi}{4}-\frac{1}{2}\arccos(x)+\frac{1}{2}x\sqrt{1-x^2}\text{.}\) We showed in part (a) that \(G(x)\) is an antiderivative of \(\sqrt{1-x^2}\text{.}\) Since \(F(x)\) is also an antiderivative of \(\sqrt{1-x^2}\text{,}\) \(F(x) = G(x)+C\) for some constant \(C\) (this is Lemma 1.3.8).Note \(G(0)=\displaystyle\int_0^0\sqrt{1-x^2}\dee{x} =0\text{,}\) so if \(F(0)=\pi\text{,}\) then \(F(x)=G(x)+\pi\text{.}\) That is,\begin{equation*} F(x) = \frac{5\pi}{4}-\frac{1}{2}\arccos(x)+\frac{1}{2}x\sqrt{1-x^2}\ . \end{equation*}
1.3.2.10.
Solution.
- The antiderivative of \(\cos x\) is \(\sin x\text{,}\) and \(\cos x\) is continuous everywhere, so \(\displaystyle\int_{-\pi}^\pi \cos x \dee{x} = \sin(\pi)-\sin(-\pi) = 0\text{.}\)
- Since \(\sec^2 x\) is discontinuous at \(x=\pm\frac{\pi}{2}\text{,}\) the Fundamental Theorem of Calculus Part 2 does not apply to \(\displaystyle\int_{-\pi}^\pi \sec^2 x \dee{x}\text{.}\)
- Since \(\frac{1}{x+1}\) is discontinuous at \(x=-1\text{,}\) the Fundamental Theorem of Calculus Part 2 does not apply to \(\displaystyle\int_{-2}^0 \frac{1}{x+1}\dee{x}\text{.}\)
1.3.2.11.
Solution.
1.3.2.12.
Solution.
1.3.2.13.
Solution.
1.3.2.14.
Solution.
The numerator describes the area of a trapezoid with base \(h\) and heights \(x\) and \(x+h\text{.}\)
\begin{align*} &=\lim_{h \to 0}\dfrac{\frac{1}{2}h(x+x+h)}{h}\\ &=\lim_{h \to 0}\left(x+\frac{1}{2}h\right)\\ &=x \end{align*}1.3.2.15.
Solution.
1.3.2.16.
Solution.
So, we know
\begin{align*} \int \log(ax)\dee{x}&= x\log(ax)-x+C \end{align*}1.3.2.17.
Solution.
So,
\begin{align*} \int x^3e^x\dee{x}&=e^x\left(x^3-3x^2+6x-6\right)+C \end{align*}1.3.2.18.
Solution.
So,
\begin{align*} \int \frac{1}{\sqrt{x^2+a^2}}\dee{x} &= \log\left|x+\sqrt{x^2+a^2}\right|+C \end{align*}1.3.2.19.
Solution.
So,
\begin{align*} \int \frac{x}{\sqrt{x(a+x)}}\dee{x}&=\sqrt{x(a+x)}-a\log\left(\sqrt{x}+\sqrt{a+x}\right)+C \end{align*}1.3.2.20. (✳).
Solution.
1.3.2.21. (✳).
Solution.
1.3.2.22.
Solution.
So, a reasonable first guess for the antiderivative might be
\begin{align*} \textcolor{red}{F(x)} &\textcolor{red}{\stackrel{?}{=} \arctan(5x)}.\\ \end{align*}However, because of the chain rule,
\begin{align*} \color{red}{F'(x)} &\color{red}{= \dfrac{5}{1+(5x)^2}}.\\ \end{align*}In order to “fix” the numerator, we make a second guess:
\begin{align*} \color{blue}{F(x)} &\color{blue}{= \frac{1}{5}\arctan(5x)}\\ \color{blue}{F'(x)} &\color{blue}{= \dfrac{1}{5}\left(\dfrac{5}{1+(5x)^2}\right) = \dfrac{1}{1+25x^2}}\\ \mbox{So,}\qquad \displaystyle\int \dfrac{1}{1+25x^2}\dee{x}&=\frac{1}{5}\arctan(5x)+C. \end{align*}1.3.2.23.
Solution.
At this point, we might guess that our antiderivative is something like \(F(x) = \arcsin\left(\dfrac{x}{\sqrt{2}}\right)\text{.}\) To explore this possibility, we can differentiate, and see what we get.
\begin{align*} \diff{}{x}\left\{\arcsin\left(\dfrac{x}{\sqrt{2}}\right)\right\}&=\frac{1}{\sqrt{2}}\cdot\dfrac{1}{\sqrt{1-\left(\dfrac{x}{\sqrt{2}}\right)^2}}\\ \end{align*}This is exactly what we want! So,
\begin{align*} \int \dfrac{1}{\sqrt{2-x^2}}\dee{x}&=\arcsin\left(\dfrac{x}{\sqrt2}\right)+C \end{align*}1.3.2.24.
Solution.
1.3.2.25.
Solution.
- Solution 1: This might not obviously look like the derivative of anything familiar, but it does look like half of a familiar trig identity: \(2\sin x \cos x = \sin(2x)\text{.}\)\begin{align*} \int 3 \sin x \cos x \dee{x}&=\int \frac{3}{2}\cdot 2\sin x \cos x \dee{x}\\ &=\int \frac{3}{2} \sin(2x)\dee{x}\\ \end{align*}
So, we might guess that the antiderivative is something like \(-\cos(2x)\text{.}\) We only need to figure out the constants.
\begin{align*} \diff{}{x}\{-\cos(2x)\}&=2\sin(2x)\\ \mbox{So,}\qquad \diff{}{x}\left\{-\frac{3}{4}\cos(2x)\right\}&=\frac{3}{2}\sin(2x)\\ \mbox{Therefore,}\qquad \int 3\sin x \cos x\dee{x}&=-\frac{3}{4}\cos(2x)+C \end{align*} - Solution 2: You might notice that the integrand looks like it came from the chain rule, since \(\cos x\) is the derivative of \(\sin x\text{.}\) Using this observation, we can work out the antideriative:\begin{align*} \diff{}{x}\left\{\sin^2 x\right\}&=2\sin x \cos x\\ \diff{}{x}\left\{\frac{3}{2}\sin^2 x\right\}&=3\sin x \cos x\\ \mbox{So,}\qquad \int 3\sin x \cos x \dee{x}&=\frac{3}{2}\sin^2 x+C \end{align*}
1.3.2.26.
Solution.
For the remaining integral, we might guess something like \(F(x) = \sin(2x)\text{.}\) Let’s figure out the appropriate constant:
\begin{align*} \diff{}{x}\left\{ \sin(2x) \right\}&=2\cos(2x)\\ \diff{}{x}\left\{ \frac{1}{4}\sin(2x) \right\}&=\frac{1}{2}\cos(2x)\\ \mbox{So,}\qquad \int \frac{1}{2}\cos(2x)\dee{x} &=\frac{1}{4}\sin(2x)+C\\ \mbox{Therefore,}\qquad \int \cos^2 x \dee{x}&=\frac{1}{2}x+\frac{1}{4}\sin(2x)+C \end{align*}1.3.2.27. (✳).
Solution.
1.3.2.28. (✳).
Solution.
1.3.2.29. (✳).
Solution.
1.3.2.30. (✳).
Solution.
1.3.2.31. (✳).
Solution.
1.3.2.32. (✳).
Solution.
1.3.2.33. (✳).
Solution.
1.3.2.34. (✳).
Solution.
1.3.2.35. (✳).
Solution.
1.3.2.36. (✳).
Solution.
1.3.2.37. (✳).
Solution.
1.3.2.38. (✳).
Solution.
1.3.2.39. (✳).
Solution.
1.3.2.40. (✳).
Solution.
1.3.2.41. (✳).
Solution.
1.3.2.42. (✳).
Solution.
1.3.2.43. (✳).
Solution.
1.3.2.44. (✳).
Solution.
1.3.2.45. (✳).
Solution.
1.3.2.46. (✳).
Solution.
1.3.2.47. (✳).
Solution.
1.3.2.48.
Solution.
-
\(\mathbf{F(x),\,x \ge 0}\text{:}\) We learned quite a lot last semester about curve sketching. We can use those techniques here. We have to be quite careful about the sign of \(x\text{,}\) though. We can only directly apply the Fundamental Theorem of Calculus Part 1 (as it’s written in your text) when \(x\ge 0\text{.}\) So first, let’s graph the right-hand portion. Notice \(f(x)\) has even symmetry--so, if we know one half of \(F(x)\text{,}\) we should be able to figure out the other half with relative ease.
- \(F(0)=\displaystyle\int_0^0 f(t)\dee{t}=0\) (so, \(F(x)\) passes through the origin)
-
Using the Fundamental Theorem of Calculus Part 1, \(F'(x) \gt 0\) when \(0 \lt x \lt 1\) and when \(3 \lt x \lt 5\text{;}\) \(F'(x) \lt 0\) when \(1 \lt x \lt 3\text{.}\) So, \(F(x)\) is decreasing from 1 to 3, and increasing from 0 to 1 and also from 3 to 5. That gives us a skeleton to work with.We get the relative sizes of the maxes and mins by eyeballing the area under \(y=f(t)\text{.}\) The first lobe (from \(x=0\) to \(x=1\) has a small positive area, so \(F(1)\) is a small positive number. The next lobe (from \(x=1\) to \(x=3\)) has a larger absolute area than the first, so \(F(3)\) is negative. Indeed, the second lobe seems to have more than twice the area of the first, so \(|F(3)|\) should be larger than \(F(1)\text{.}\) The third lobe is larger still, and even after subtracting the area of the second lobe it looks much larger than the first or second lobe, so \(|F(3)| \lt F(5)\text{.}\)
-
We can use \(F''(x)\) to get the concavity of \(F(x)\text{.}\) Note \(F''(x)=f'(x)\text{.}\) We observe \(f(x)\) is decreasing on (roughly) \((0,2.5)\) and \((4,5)\text{,}\) so \(F(x)\) is concave down on those intervals. Further, \(f(x)\) is increasing on (roughly) \((2.5,4)\text{,}\) so \(F(x)\) is concave up there, and has inflection points at about \(x=2.5\) and \(x=4\text{.}\)In the sketch above, closed dots are extrema, and open dots are inflection points.
-
\(\mathbf{F(x),\,x \lt 0}\text{:}\) Now we can consider the left half of the graph. If you stare at it long enough, you might convince yourself that \(F(x)\) is an odd function. We can also show this with the following calculation:\begin{align*} F(-x)&=\int_0^{-x} f(t)\dee{t} =\int_x^0 f(t)\dee{t}\\ &\hskip0.5in\text{as in Example }\knowl{./knowl/xref/eg_lefthalfevenfunction.html}{\text{1.2.10}}\text{, since $f(t)$ is even,}\\ &=-\int_0^x f(t)\dee{t}\\ &=-F(x) \end{align*}Knowing that \(F(x)\) is odd allows us to finish our sketch.
1.3.2.49. (✳).
Solution.
So, the equation of the tangent line is
\begin{align*} y = -3(x+1)\ . \end{align*}1.3.2.50.
Solution.
1.3.2.51.
Solution.
- When \(x=3\text{,}\)\begin{align*} F(3)&=\displaystyle\int_0^3 3^3 \sin(t)\dee{t}=27\int_0^3 \sin t \dee{t}\\ \end{align*}
Using the Fundamental Theorem of Calculus Part 2,
\begin{align*} &=27\left[-\cos t\right]_{t=0}^{t=3} =27\left[-\cos 3 - (-\cos 0)\right]\\ &=27(1-\cos 3 ) \end{align*} - Since the integration is with respect to \(t\text{,}\) the \(x^3\) term can be moved outside the integral. That is: for the purposes of the integral, \(x^3\) is a constant (although for the purposes of the derivative, it certainly is not).\begin{align*} F(x)&=\displaystyle\int_0^x x^3 \sin(t)\dee{t} = x^3 \int_0^x \sin(t)\dee{t}\\ \end{align*}
Using the product rule and the Fundamental Theorem of Calculus Part 1,
\begin{align*} F'(x)&=x^3\cdot \sin(x) + 3x^2 \int_0^x \sin(t)\dee{t}\\ &=x^3\sin(x)+3x^2\left[-\cos(t)\right]_{t=0}^{t=x}\\ &=x^3\sin(x)+3x^2[-\cos(x)-(-\cos(0))]\\ &=x^3\sin (x) + 3x^2[1-\cos (x)] \end{align*}
1.3.2.52.
Solution.
By the definition of \(G(x)\text{,}\)
\begin{align*} G(x)&=G(-x) \end{align*}
1.4 Substitution
1.4.2 Exercises
1.4.2.1.
Solution.
1.4.2.2.
Solution.
Problem: Evaluate \(\displaystyle\int (2x+1)^2 \dee{x}\text{.}\)Work: We use the substitution \(u=2x+1\text{.}\) Then \(\dee{u}=2\dee{x}\text{,}\) so \(\dee{x} = \frac{1}{2}\dee{u}\text{:}\)\begin{align*} \int (2x+1)^2 \dee{x}&=\int u^2\cdot \frac{1}{2}\dee{u}\\ &=\frac{1}{6}u^3+C\\ &=\frac{1}{6}\left(2x+1\right)^3+C \end{align*}
1.4.2.3.
Solution.
Problem: Evaluate \(\displaystyle\int_{1}^{\pi} \dfrac{\log(\log t)}{t}\dee{t}\text{.}\)Work: We use the substitution \(u=\log t\text{,}\) so \(\dee{u}=\frac{1}{t}\dee{t}\text{.}\) When \(t=1\text{,}\) we have \(u=\log 1 =0\) and when \(t=\pi\text{,}\) we have \(u=\log(\pi)\text{.}\) Then:\begin{align*} \int_{1}^{\pi} \dfrac{\cos(\log t)}{t}\dee{t}&=\int_{\log 1}^{\log(\pi)}\cos(u) \dee{u}\\ &=\int_{0}^{\log(\pi)}\cos(u) \dee{u}\\ &=\sin(\log(\pi))-\sin(0)=\sin(\log(\pi)) . \end{align*}
1.4.2.4.
Solution.
Problem: Evaluate \(\displaystyle\int_{0}^{\pi/4} x\tan (x^2) \dee{x}\text{.}\)Work: We begin with the substitution \(u=x^2\text{,}\) \(\dee{u} = 2x\dee{x}\text{:}\) If \(u=x^2\text{,}\) then \(\diff{u}{x} = 2x\text{,}\) so indeed \(\dee{u}=2x\dee{x}\text{.}\)\begin{align*} \int_{0}^{\pi/4} x\tan (x^2) \dee{x}&= \int_{0}^{\pi/4} \frac{1}{2}\tan(x^2)\cdot 2x\dee{x}&\color{red}{\text{algebra}}\\ &=\int_{0}^{\pi^2/16} \frac{1}{2}\tan u\dee{u} \end{align*}Note that every piece was changed from \(x\) to \(u\text{:}\) integrand, differential, limits. So\begin{align*} \int_{0}^{\pi/4} x\tan (x^2) \dee{x} &=\frac{1}{2}\int_{0}^{\pi^2/16} \dfrac{\sin u}{\cos u}\dee{u} \end{align*}since \(\tan u = \frac{\sin u}{\cos u}\text{.}\) Now we use the substitution \(v=\cos u\text{,}\) \(\dee{v}=-\sin u \dee{u}\text{:}\)\begin{align*} \frac{1}{2}\int_{0}^{\pi^2/16} \dfrac{\sin u}{\cos u}\dee{u} &=\frac{1}{2}\int_{\cos 0}^{\cos(\pi^2/16)} -\dfrac{1}{v}\dee{v} \end{align*}Note that every piece was changed from \(u\) to \(v\text{:}\) integrand, differential, limits. So\begin{align*} \int_{0}^{\pi/4} x\tan (x^2) \dee{x} &=-\frac{1}{2}\int_{1}^{\cos(\pi^2/16)} \dfrac{1}{v}\dee{v}\\ &\hskip1in \text{since }\color{red}{\cos(0)=1}\\ &=-\frac{1}{2}\bigg[\log|v|\bigg]_{1}^{\cos(\pi^2/16)}\\ &\hskip1in\color{red}{\text{FTC Part 2}}\\ &=-\frac{1}{2}\left(\log\left(\cos(\pi^2/16)\right)-\log(1)\right)\\ &=-\frac{1}{2}\log\left(\cos(\pi^2/16)\right)\\ &\hskip1in \text{since }\color{red}{\log(1)=0} \end{align*}
1.4.2.5. (✳).
Solution.
So,
\begin{align*} \int_{x=0}^{x=\pi/2} f(\sin x)\,\dee{x} &= \int_{u=0}^{u=1} f(u)\,\frac{\dee{u}}{\sqrt{1-u^2}} \end{align*}1.4.2.6.
Solution.
1.4.2.7. (✳).
Solution.
1.4.2.8. (✳).
Solution.
1.4.2.9. (✳).
Solution.
1.4.2.10. (✳).
Solution.
1.4.2.11. (✳).
Solution.
1.4.2.12. (✳).
Solution.
1.4.2.13. (✳).
Solution.
1.4.2.14. (✳).
Solution.
1.4.2.15.
Solution.
- Solution 1: If we let \(\textcolor{red}{u=\sqrt{\log x}}\text{,}\) then \(\textcolor{blue}{\dee{u}=\dfrac{1}{2x\sqrt{\log x}}\dee{x}}\text{,}\) and:\begin{align*} \int \dfrac{e^{\textcolor{red}{\sqrt{\log x}}}}{\textcolor{blue}{2x\sqrt{\log x}}}\ \textcolor{blue}{\dee{x}}&=\int e^{\textcolor{red}u}\ \textcolor{blue}{\dee{u}}=e^u+C=e^{\sqrt{\log x}}+C \end{align*}
- Solution 2: In Solution 1, we made a pretty slick choice. We might have tried to work with something a little less convenient. For example, it’s not unnatural to think that \(\textcolor{red}{u=\log x}\text{,}\) \(\textcolor{blue}{\dee{u}=\dfrac{1}{x}\dee{x}}\) would be a good choice. In that case:\begin{align*} \int \dfrac{e^{\sqrt{\textcolor{red}{\log x}}}}{2\textcolor{blue}{x}\sqrt{\textcolor{red}{\log x}}}\ \textcolor{blue}{\dee{x}}&= \int \frac{e^{\sqrt{\textcolor{red}u}}}{2\sqrt{\textcolor{red}u}}\textcolor{blue}{\dee{u}}\\ \end{align*}
Now, we should be able to see that \(\textcolor{orange!40!black}{w=\sqrt{u}}\text{,}\) \(\textcolor{purple}{\dee{w} = \dfrac{1}{2\sqrt{u}}\dee{u}}\) is a good choice:
\begin{align*} \int\frac{e^{\textcolor{orange!40!black}{\sqrt{u}}}}{\textcolor{purple}{2\sqrt{u}}}\ \textcolor{purple}{\dee{u}} &=\int e^{\textcolor{orange!40!black}{w}}\ \textcolor{purple}{\dee{w}}\\ &=e^{\sqrt{u}}+C\\ &=e^{\sqrt{\log x}}+C \end{align*}
1.4.2.16. (✳).
Solution.
- The straightforward method: We use the substitution \(\textcolor{red}{u=x^2}\text{,}\) for which \(\textcolor{blue}{\dee{u}=2x\,\dee{x}}\text{,}\) and note that \(u=4\) for both \(x=2\) and \(x=-2\text{:}\)\begin{equation*} \int_{-2}^2 xe^{x^2}\,\dee{x} =\int_{-2}^2 \frac{1}{2}e^{\textcolor{red}{x^2}}\,\textcolor{blue}{2x\dee{x}} = \int_4^4 \frac{1}{2}e^{\textcolor{red}{u}}\,\textcolor{blue}{\dee{u}} =0 \end{equation*}
- The slightly sneaky method: We note that \(\displaystyle\diff{}{x} \left\{e^{x^2} \right\}= 2x\, e^{x^2}\text{,}\) so that \(\dfrac{1}{2} e^{x^2}\) is a antiderivative for the integrand \(x e^{x^2}\text{.}\) So\begin{gather*} \int_{-2}^2 xe^{x^2}\,\dee{x} = \bigg[\frac{1}{2}e^{x^2}\bigg]_{-2}^2 =\frac{1}{2}e^4-\frac{1}{2}e^4=0 \end{gather*}
- The really sneaky method: The integrand \(f(x) = x e^{x^2}\) is an odd function (meaning that \(f(-x)=-f(x)\)). So by Theorem 1.2.12 every integral of the form \(\int_{-a}^a x e^{x^2}\,\dee{x}\) is zero.
1.4.2.17. (✳).
Solution.
1.4.2.18.
Solution.
The numerator now is \(u^2\text{,}\) and looking at our substitution, we see \(\textcolor{red}{u^2=w-1}\text{:}\)
\begin{align*} &=\frac{1}{2}\int_{1}^2 \dfrac{\textcolor{red}{w-1}}{\textcolor{red}{w}}\ \textcolor{blue}{\dee{w}}\\ &=\frac{1}{2}\int_{1}^2 \left(1 - \frac{1}{w}\right)\dee{w}\\ &=\frac{1}{2}\left[w - \log|w|\right]_{w=1}^{w=2}\\ &=\frac{1}{2}\left(2-\log 2 - 1\right)=\frac{1}{2}-\frac{1}{2}\log 2 \end{align*}1.4.2.19.
Solution.
In Example 1.4.17, we learned \(\int \tan \theta \dee{\theta} = \log |\sec \theta|+C\)
\begin{align*} &=\int \textcolor{red}{u}\ \textcolor{blue}{\dee{u}} -\log|\sec \theta|+C\\ &=\frac{1}{2}u^2 -\log|\sec \theta|+C\\ &=\frac{1}{2}\tan^2\theta -\log|\sec \theta|+C \end{align*}1.4.2.20.
Solution.
1.4.2.21.
Solution.
The left integral is tough to solve with substitution, but luckily we don’t have to--it’s the area of a quarter of a circle of radius 1.
\begin{align*} &=\frac{\pi}{4}+\int_1^0 \sqrt{\textcolor{red}{u}}\ \textcolor{blue}{\dee{u}}\\ &=\frac{\pi}{4}+\left[\frac{2}{3}u^{3/2}\right]_{u=1}^{u=0}\\ &=\frac{\pi}{4} + 0 - \frac{2}{3} = \frac{\pi}{4}-\frac{2}{3} \end{align*}1.4.2.22.
Solution.
- Solution 1: We often find it useful to take “inside” functions as our substitutions, so let’s try \(\textcolor{red}{u=\cos x}\text{,}\) \(\textcolor{blue}{\dee{u} = -\sin x\dee{x}}\text{.}\) In order to dig up a sine, we use the identity \(\tan x = \dfrac{\sin x}{\cos x}\text{:}\)\begin{align*} \int\tan x \cdot \log\left(\textcolor{red}{\cos x}\right) \dee{x}&= -\int\frac{\textcolor{blue}{-\sin x}}{\textcolor{red}{\cos x}} \cdot \log\left(\textcolor{red}{\cos x}\right) \textcolor{blue}{\dee{x}}\\ &=-\int\frac{1}{\textcolor{red}u}\log(\textcolor{red}u)\textcolor{blue}{\dee{u}}\\ \end{align*}
Now, it is convenient to let \(\textcolor{orange!40!black}{w=\log u}\text{,}\) \(\textcolor{purple}{\dee{w}=\frac{1}{u}\dee{u}}\text{:}\)
\begin{align*} -\int\textcolor{purple}{\frac{1}{u}}\textcolor{orange!40!black}{\log(u)}\textcolor{purple}{\dee{u}} &=-\int \textcolor{orange!40!black}{w}\ \textcolor{purple}{\dee{w}}\\ &=-\frac{1}{2}w^2+C\\ &=-\frac{1}{2}\left(\log u\right)^2+C\\ &=-\frac{1}{2}\left(\log (\cos x)\right)^2+C \end{align*} - Solution 2: We might guess that it’s useful to have \(\textcolor{red}{u=\log(\cos x)}\text{,}\) \(\textcolor{blue}{\dee{u}=\dfrac{-\sin x}{\cos x}\dee{x} = -\tan x\dee{x}}\text{:}\)\begin{align*} \int\tan x \cdot \textcolor{red}{\log\left(\cos x\right)} \dee{x}&= -\int\textcolor{blue}{-\tan x} \cdot \textcolor{red}{\log\left(\cos x\right) }\textcolor{blue}{\dee{x}}\\ &=-\int \textcolor{red}{u}\ \textcolor{blue}{\dee{u}}\\ &=-\frac{1}{2}u^2+C\\ &=-\frac{1}{2}\left(\log(\cos x)\right)^2+C \end{align*}
1.4.2.23. (✳).
Solution.
1.4.2.24. (✳).
Solution.
1.4.2.25.
Solution.
Since the Riemann sums are exactly the same,
\begin{align*} \color{red}{\int_a^b 2f(2x)\dee{x}}&= \color{blue}{\int_{2a}^{2b} f(x)\dee{x}} \end{align*}
1.5 Area between curves
1.5.2 Exercises
1.5.2.1.
Solution.
- \(x = 0\text{:}\) The distance from \(\cos 0\) to \(\sin 0\) is 1, so our first rectangle has height 1.
- \(x = \frac{\pi}{4}\text{:}\) The distance from \(\cos \frac{\pi}{4}\) to \(\sin \frac{\pi}{4}\) is 0, so our second rectangle has height 0.
- \(x = \frac{\pi}{2}\text{:}\) The distance from \(\cos \frac{\pi}{2}\) to \(\sin \frac{\pi}{2}\) is 1, so our third rectangle has height 1.
- \(x = \frac{3\pi}{4}\text{:}\) The distance from \(\cos \frac{3\pi}{4}\) to \(\sin \frac{3\pi}{4}\) is \(\sin(3\pi/4)-\cos(3\pi/4) =\frac{1}{\sqrt{2}}-\left(-\frac{1}{\sqrt{2}}\right)=\sqrt{2} \text{,}\) so our fourth rectangle has height \(\sqrt{2}\text{.}\)
1.5.2.2.
Solution.
-
We are finding the area in the interval from \(x=0\) to \(x=\frac{\pi}{2}\text{.}\) Since we’re taking \(n=5\) rectangles, our rectangles cover the following intervals:\begin{equation*} \left[0,\frac{\pi}{10}\right],\quad \left[\frac{\pi}{10},\frac{\pi}{5}\right],\quad \left[\frac{\pi}{5},\frac{3\pi}{10}\right],\quad \left[\frac{3\pi}{10},\frac{2\pi}{5}\right],\quad \left[\frac{2\pi}{5},\frac{\pi}{2}\right]. \end{equation*}
-
We are finding the area in the interval from \(y=0\) to \(y=\frac{\pi}{2}\text{.}\) (In general, when we switch from horizontal rectangles to vertical, the limits of integration will change--it’s only coincidence that they are the same in this example.) Since we’re taking \(n=5\) rectangles, these rectangles cover the following intervals of the \(y\)-axis:\begin{equation*} \left[0,\frac{\pi}{10}\right],\quad \left[\frac{\pi}{10},\frac{\pi}{5}\right],\quad \left[\frac{\pi}{5},\frac{3\pi}{10}\right],\quad \left[\frac{3\pi}{10},\frac{2\pi}{5}\right],\quad \left[\frac{2\pi}{5},\frac{\pi}{2}\right]. \end{equation*}The question doesn’t specify which endpoints we’re using. Let’s use upper endpoints, to match part (a).
1.5.2.3. (✳).
Solution.
1.5.2.4. (✳).
Solution.
1.5.2.5. (✳).
Solution.
1.5.2.6. (✳).
Solution.
1.5.2.7. (✳).
Solution.
1.5.2.8. (✳).
Solution.
1.5.2.9. (✳).
Solution.
3
1.5.2.10. (✳).
Solution.
4
1.5.2.11. (✳).
Solution.
1.5.2.12. (✳).
Solution.
- The curve intercepts the \(y\)-axis when \(y=0\) and \(y=-1\text{.}\)
- The \(x\)-values of the curve are negative when \(-1 \lt y \lt 0\text{,}\) and positive elsewhere.
1.5.2.13.
Solution.
1.5.2.14. (✳).
Solution.
\(x\) | \(4+2\pi-2x\) | \(4+\pi\sin(x)\) | match? |
\(\frac{\pi}{2}\) | \(4+\pi\) | \(4+\pi \) | \(\checkmark\) |
\(\pi\) | \(4\) | \(4\) | \(\checkmark\) |
\(\frac{3\pi}{2}\) | \(4-\pi\) | \(4-\pi \) | \(\checkmark\) |
- When \(\frac{\pi}{2} \le x \le \pi\text{,}\) the top of the strip is at \(y = 4 + \pi \sin x\) and the bottom of the strip is at \(y = 4 + 2\pi - 2x\text{.}\) So the strip has height \(\big[(4 + \pi \sin x)-(4 + 2\pi - 2x)\big]\) and width \(\dee{x}\text{,}\) and hence area \(\big[(4 + \pi \sin x)-(4 + 2\pi - 2x)\big]\dee{x}\text{.}\)
- When \(\pi \le x \le \frac{3\pi}{2}\text{,}\) the top of the strip is at \(y = 4 + 2\pi - 2x\) and the bottom of the strip is at \(y = 4 + \pi \sin x\text{.}\) So the strip has height \(\big[(4 + 2\pi - 2x)-(4 + \pi \sin x)\big]\) and width \(\dee{x}\text{,}\) and hence area \(\big[(4 + 2\pi - 2x)-(4 + \pi \sin x)\big]\dee{x}\text{.}\)
1.5.2.15. (✳).
Solution.
1.5.2.16. (✳).
Solution.
We only care about values of \(x\) in \([0,4]\text{,}\) so \(x\) is nonnegative. Then \(h(x)\) is positive when:
\begin{align*} 3& \gt \sqrt{25-x^2}\\ 9& \gt 25-x^2\\ x^2 & \gt 16\\ x& \gt 4 \end{align*}1.5.2.17.
Solution.
- The desired area is \(A_3-(A_1+A_2)\text{.}\)
- \(A_1\) is the area of right a triangle with base 1 and height 1, so \(A_1 = \frac{1}{2}\text{.}\)
- \(A_2\) is the area of a quarter circle of radius 1, so \(A_2=\frac{\pi}{4}\text{.}\)
- \(A_3\) is the area of an eighth of a circle of radius 3, so \(A_3 = \frac{9\pi}{8}\)
1.5.2.18.
Solution.
\(x\) | \(x(x^2-4)\) | \(x-2\) | top function: |
0 | 0 | \(-2\) | \(x(x^2-4)\) |
1 | -3 | \(-1\) | \(x-2\) |
After some taxing but rudimentary algebra:
\begin{align*} &=\left(8\sqrt{2}\right)+\left(4\sqrt{2}-\frac{13}{4}\right)=12\sqrt{2}-\frac{13}{4} \end{align*}
1.6 Volumes
1.6.2 Exercises
1.6.2.1.
Solution.
1.6.2.2.
Solution.
1.6.2.3.
Solution.
-
Washers when \(\mathbf{1 \lt y \le 6}\text{:}\) If \(y \gt 1\text{,}\) then our washer has inner radius \(2+\frac{2}{3}y\text{,}\) outer radius \(6-\frac{2}{3}y\text{,}\) and height \(\dee{y}\text{.}\)
-
Washers when \(\mathbf{0\le y \lt 1}\text{:}\) When \(0 \le y \lt 1\text{,}\) we have a “double washer,” two concentric rings corresponding to the two “humps” in the function. The inner washer has inner radius \(r_1=y\) and outer radius \(R_1=2-y\text{.}\) The outer washer has inner radius \(r_2=2+\frac{2}{3}y\) and outer radius \(R_2=6-\frac{2}{3}y\text{.}\) The thickness of the washers is \(\dee{y}\text{.}\)
1.6.2.4. (✳).
Solution.
- We use thin horizontal strips of width \(\dee{y}\) as in the figure above.
- When we rotate about the line \(x=3\text{,}\) each strip sweeps out a thin washer
- whose inner radius is \(r_{in}=3-\sqrt{y}\text{,}\) and
- whose outer radius is \(r_{out}=3-(y-2)=5-y\) when \(y\ge 1\) (see the red strip in the figure on the right above), and whose outer radius is \(r_{out}= 3-(-\sqrt{y})=3+\sqrt{y}\) when \(y\le 1\) (see the blue strip in the figure on the right above) and
- whose thickness is \(\dee{y}\) and hence
- whose volume is \(\pi(r_{out}^2 - r_{in}^2)\dee{y} = \pi\big[\big(5-y\big)^2-\big(3-\sqrt{y}\big)^2\big]\dee{y}\) when \(y\ge 1\) and whose volume is \(\pi(r_{out}^2 - r_{in}^2)\dee{y} =\pi\big[\big(3+\sqrt{y}\big)^2-\big(3-\sqrt{y}\big)^2\big]\dee{y}\) when \(y\le 1\) and
- As our bottommost strip is at \(y=0\) and our topmost strip is at \(y=4\text{,}\) the total volume is\begin{align*} &\int _0^1 \pi\big[\big(3+\sqrt{y}\big)^2-\big(3-\sqrt{y}\big)^2\big]\dee{y}\\ &\hskip0.5in+\int _ 1^4 \pi\big[\big(5-y\big)^2-\big(3-\sqrt{y}\big)^2\big]\dee{y} \end{align*}
1.6.2.5. (✳).
Solution.
- inner radius \((1-x^2)-(-1)=2-x^2\text{,}\)
- outer radius \((4-4x^2)-(-1)=5-4x^2\) and
- thickness \(\dee{x}\ \text{;}\) so, it has
- cross sectional area \(\pi\big[{(5-4x^2)}^2-{(2-x^2)}^2\big]\) and
- volume \(\pi\big[{(5-4x^2)}^2-({2-x^2)}^2\big]\,\dee{x}\text{.}\)
- We use thin horizontal strips of height \(\dee{y}\) as in the figure above.
- When we rotate about the line \(x=5\text{,}\) each strip sweeps out a thin washer
- whose inner radius is \(r_{in}=5-\sqrt{y+1}\text{,}\) and
- whose outer radius is \(r_{out}= 5-(-\sqrt{y+1})=5+\sqrt{y+1}\) and
- whose thickness is \(\dee{y}\) and hence
- whose volume is \(\pi(r_{out}^2 - r_{in}^2)\,\dee{y} = \pi\big[\big(5+\sqrt{y+1}\big)^2-\big(5-\sqrt{y+1}\big)^2\big]\,\dee{y}\)
- As our topmost strip is at \(y=0\) and our bottommost strip is at \(y=-1\) (when \(x=0\)), the total volume is\begin{gather*} \int _{-1}^0 \pi\left[\big(5+\sqrt{y+1}\big)^2-\big(5-\sqrt{y+1}\big)^2\right]\,\dee{y} \end{gather*}
1.6.2.6. (✳).
Solution.
- inner radius \(x^2+1\text{,}\)
- outer radius \(9-x^2\) and
- thickness \(\dee{x}\) and hence of
- cross sectional area \(\pi\big[{(9-x^2)}^2-{(x^2+1)}^2\big]\) and
- volume \(\pi\big[{(9-x^2)}^2-{(x^2+1)}^2\big]\,\dee{x}\text{.}\)
1.6.2.7.
Solution.
1.6.2.8. (✳).
Solution.
1.6.2.9. (✳).
Solution.
1.6.2.10. (✳).
Solution.
1.6.2.11. (✳).
Solution.
1.6.2.12. (✳).
Solution.
- We cut \(R\) into thin horizontal strips of height \(\dee{y}\) as in the figure on the right above.
- When we rotate \(R\) about the \(y\)--axis, i.e. about the line \(x=0\text{,}\) each strip sweeps out a thin washer
- whose inner radius is \(r_{in}= e^y\) and outer radius is \(r_{out}=2\text{,}\) and
- whose thickness is \(\dee{y}\) and hence
- whose volume \(\pi(r_{out}^2 - r_{in}^2)\dee{y} = \pi\big(4-e^{2y}\big)\dee{y}\text{.}\)
- As our bottommost strip is at \(y=0\) and our topmost strip is at \(y=\log 2\) (since at the top \(x=2\) and \(x=e^y\)), the total\begin{align*} \text{Volume} &= \int _0^{\log 2} \pi\big(4-e^{2y}\big)\dee{y} =\pi\big[4y -e^{2y}/2\big]_0^{\log 2}\\ &=\pi\Big[4\log 2 - 2 +\frac{1}{2}\Big] =\pi\Big[4\log 2 - \frac{3}{2}\Big] \end{align*}Using a calculator, we see this is approximately \(3.998\text{.}\)
1.6.2.13. (✳).
Solution.
- We cut the specified region into thin vertical strips of width \(\dee{x}\) as in the figure above.
- When we rotate about the line \(y=-\pi^2\text{,}\) each strip sweeps out a thin washer
- whose inner radius is \(r_{in}= (x^2 - {\pi^2} ) - ( {-} {\pi^2} )=x^2\) and outer radius is \(r_{out}=\cos(\frac x2) - ( {-} {\pi^2}) =\cos(\frac x2) +\pi^2 \text{,}\) and
- whose thickness is \(\dee{x}\) and hence
- whose volume \(\pi(r_{out}^2 - r_{in}^2)\dee{x} = \pi\big( {(\cos(\frac x2) +\pi^2)}^2 - {(x^2)}^2\big)\dee{x}\text{.}\)
- As our leftmost strip is at \(x=-\pi\) and our rightmost strip is at \(x=\pi\text{,}\)
Because the integrand is even,
\begin{align*} & = 2\pi \int_0^{\pi} \left(\frac{1+\cos(x)}{2} +2{\pi^2}\cos (\tfrac x2) +{\pi^4} -x^4\right)\,\dee{x}\\ & = {2\pi \left[ \frac{1}{2}x + \frac{1}{2}\sin(x) + 4{\pi^2}\sin (\tfrac x2) +{\pi^4} x -\frac{1}{5}x^5 \right]} _0^\pi\\ & = 2\pi \left[\frac{\pi}{2} + 0 + 4{\pi^2} +{\pi^5} - \frac{\pi^5}{5} \right]\\ & = {\pi^2} + 8\pi^3 + \frac{8\pi^6}{5} \end{align*}1.6.2.14. (✳).
Solution.
- We are told that the pancake at height \(x\) is a square of side \(\frac{2}{1+x}\) and so
- has cross-sectional area \(\big(\frac{2}{1+x}\big)^2\) and thickness \(\dee{x}\) and hence
- has volume \(\big(\frac{2}{1+x}\big)^2\dee{x}\text{.}\)
1.6.2.15. (✳).
Solution.
1.6.2.16. (✳).
Solution.
1.6.2.17.
Solution.
So, our absolute error is:
\begin{align*} &\left|\frac{4\pi}{3}\left(6378.137\right)^2\left(6356.752\right) - \dfrac{4}{3}\pi\left(6378.137\right)^3 \right|\\ =&\frac{4\pi}{3}\left(6378.137\right)^2\big|6356.752 - 6378.137\big|\\ =&\frac{4\pi}{3}\left(6378.137\right)^2(21.385)\\ \approx& 3.64 \times 10^{9} \mathrm{km}^3\\ \end{align*}And our relative error is:
\begin{align*} \frac{\mbox{abs error}}{\mbox{actual value}}&=\frac{\frac{4\pi}{3}\left(6378.137\right)^2\big|6356.752 - 6378.137\big|}{\frac{4\pi}{3}\left(6378.137\right)^2\left(6356.752\right)}\\ &=\frac{\big|6356.752 - 6378.137\big|}{6356.752}\\ &=\frac{6378.137}{6356.752}-1\\ &\approx 0.00336 \end{align*}1.6.2.18. (✳).
Solution.
- We cut \(R\) into thin vertical strips of width \(\dee{x}\) like the red strip in the figure above.
- When we rotate \(R\) about the horizontal line \(y=5\text{,}\) each strip sweeps out a thin washer
- whose inner radius is \(r_{in}=5-[4-(x-1)^2]=1+(x-1)^2\text{,}\) and
- whose outer radius is \(r_{out}= 5-[x+1]=4-x\) and
- whose thickness is \(\dee{x}\) and hence
- whose volume is \(\pi\big[r_{out}^2-r_{in}^2\big]\,\dee{x} =\pi\big[{\big(4-x\big)}^2-{\big(1+(x-1)^2\big)}^2\big]\,\dee{x}\)
- As our leftmost strip is at \(x=-1\) and our rightmost strip is at \(x=2\text{,}\) the total\begin{align*} {\rm Volume} &= \pi\int_{-1}^2 \big[{\big(4-x\big)}^2-{\big(1+(x-1)^2\big)}^2\big]\,\dee{x} \end{align*}
1.6.2.19. (✳).
Solution.
Now, we calculate:
\begin{align*} \text{Area} &= 2\int_0^1 \big[y-\big(1-\sqrt{1-y^2}\big)\big]\dee{y}\\ &=2\left\{ \int_0^1 y-1\dee{y} + \int_0^1 \sqrt{1-y^2}\dee{y} \right\}\\ &=2\Big\{\Big[\frac{y^2}{2}-y\Big]_0^1 +\int_0^1\sqrt{1-y^2}\dee{y}\Big\}\\ &=\frac{\pi}{2}-1 \end{align*}- We cut \(\cR\) into thin horizontal strips of width \(\dee{y}\) like the blue strip in the figure above.
- When we rotate \(\cR\) about the \(y\)--axis, each strip sweeps out a thin washer
- whose inner radius is \(r_{in}=1-\sqrt{1-y^2}\text{,}\) and
- whose outer radius is \(r_{out}= \sqrt{1-(y-1)^2}\) and
- whose thickness is \(\dee{y}\) and hence
- whose volume is\begin{align*} &\pi\big[{\big(\sqrt{1-(y-1)^2}\big)}^2-{\big(1-\sqrt{1-y^2}\,\big)}^2\big]\,\dee{y}\\ =&\pi \big[ 1 - (y-1)^2 -1 + 2\sqrt{1-y^2} - (1-y^2) \big]\\ =&2\pi\big[\sqrt{1-y^2}+y-1\big]\,\dee{y} \end{align*}
- As our bottommost strip is at \(y=0\) and our topmost strip is at \(y=1\text{,}\) the total\begin{align*} {\rm Volume} &= 2\pi\int_{0}^1\big[\sqrt{1-y^2}+y-1\big]\dee{y} = 2\pi\Big[\frac{\pi}{4}+\frac{1}{2}-1\Big]\\ &=\frac{\pi^2}{2}-\pi\approx 1.793 \end{align*}Here, we again used that \(\int_{0}^1 \sqrt{1-y^2}\ \dee{y}\) is the area of a quarter circle of radius one, and we used a calculator to approximate the final answer.
1.6.2.20. (✳).
Solution.
- We cut \(\cR\) into thin horizontal strips of width \(\dee{y}\) as in the figure on the right above.
- When we rotate \(\cR\) about the \(y\)--axis, i.e. about the line \(x=0\text{,}\) each strip sweeps out a thin washer
- whose outer radius is \(r_{out}=1\text{,}\) and
- whose inner radius is \(r_{in}= \sqrt{\frac{y^2}{c^2}-1}\) when \(y\ge c\sqrt{1+0^2}=c\) (see the red strip in the figure on the right above), and whose inner radius is \(r_{in}= 0\) when \(y\le c\) (see the blue strip in the figure on the right above) and
- whose thickness is \(\dee{y}\) and hence
- whose volume is \(\pi(r_{out}^2 - r_{in}^2)\dee{y} = \pi\big(2-\frac{y^2}{c^2}\big)\dee{y}\) when \(y\ge c\) and whose volume is \(\pi(r_{out}^2 - r_{in}^2)\dee{y} = \pi\,\dee{y}\) when \(y\le c\) and
- As our bottommost strip is at \(y=0\) and our topmost strip is at \(y=\sqrt{2}\,c\) (since at the top \(x=1\) and \(y= c\sqrt{1+x^2}\)), the total\begin{align*} V_2 &= \int _c^{\sqrt{2}\,c} \pi\Big(2-\frac{y^2}{c^2}\Big)\dee{y} +\int _ 0^c \pi\,\dee{y}\\ &=\pi{\Big[2y -\frac{y^3}{3c^2}\Big]}_c^{\sqrt{2}\,c} +\pi c\\ &=\pi\,c\Big[\frac{4\sqrt{2}}{3}-\frac{5}{3} \Big]+\pi c\\ &=\frac{\pi\,c}{3}\big[4\sqrt{2}-2 \big] \end{align*}
1.6.2.21. (✳).
Solution.
- When \(\frac{\pi}{2} \le x \le \pi\text{,}\) the top of the strip is at \(y = 4 + \pi \sin x\) and the bottom of the strip is at \(y = 4 + 2\pi - 2x\text{.}\) When the strip is rotated, we get a thin washer with outer radius \(R_1(x)= 1+ 4 + \pi \sin x=5 + \pi \sin x \) and inner radius \(r_1(x) = 1+4 + 2\pi - 2x=5 + 2\pi - 2x\text{.}\)
- When \(\pi \le x \le \frac{3\pi}{2}\text{,}\) the top of the strip is at \(y = 4 + 2\pi - 2x\) and the bottom of the strip is at \(y = 4 + \pi \sin x\text{.}\) When the strip is rotated, we get a thin washer with outer radius \(R_2(x) = 1+4 + 2\pi - 2x=5 + 2\pi - 2x\) and inner radius \(r_2(x) = 1+ 4 + \pi \sin x=5 + \pi \sin x\text{.}\)
1.6.2.22.
Solution.
Since \(k\) runs from 0 to \(60,000\text{,}\) the total mass is given by
\begin{align*} \int_0^{60000} c\pi2^{-k/6000} \dee{k}&=c\pi\int_0^{60000} 2^{-k/6000} \dee{k}\\ \end{align*}To facilitate integration, we can write our exponential function in terms of \(e\text{,}\) then use the substitution \(u=-\frac{k}{6000}\log 2\text{,}\) \(\dee{u} = -\frac{1}{6000}\log 2 \dee{k}\text{.}\)
\begin{align*} &=c\pi\int_0^{60000}\left(e^{\log 2}\right)^{-k/6000}\dee{k}\\ &=c\pi\int_0^{60000}e^{-\tfrac{k}{6000}\log 2}\dee{k}\\ &=-\frac{6000c\pi}{\log 2}\int_0^{-10\log 2}e^{u}\dee{u}\\ &=\frac{6000c\pi}{\log 2}\int_{-10\log 2}^0e^{u}\dee{u}\\ &=\frac{6000c\pi}{\log 2}\left(1-\frac{1}{2^{10}}\right) \end{align*}So, we set this equal to the mass we want, and solve for \(k\text{.}\)
\begin{align*} \frac{6000c\pi}{\log 2}\left(1-\frac{1}{2^{k/6000}}\right)&=\frac{3000c\pi}{\log 2}\\ 2\left(1-\frac{1}{2^{k/6000}}\right)&=1\\ 1&=\frac{2}{2^{k/6000}}\\ 2^{k/6000}&=2^1\\ k&=6000\\ h&=6 \end{align*}
1.7 Integration by parts
1.7.2 Exercises
1.7.2.1.
Solution.
Similarly, integration by parts comes from the product rule:
\begin{align*} &&\diff{}{x}\{f(x)g(x)\}&=f'(x)g(x)+f(x)g'(x)\\ &\Leftrightarrow&\int\diff{}{x}\{f(x)g(x)\}\dee{x}&=\int f'(x)g(x)+f(x)g'(x)\dee{x}\\ &\Leftrightarrow&f(x)g(x)+C&=\int f'(x)g(x)\dee{x}+\int f(x)g'(x)\dee{x}\\ &\Leftrightarrow&\int\underbrace{ f(x)}_{u}\underbrace{g'(x)\dee{x}}_{\dee{v}}&=\underbrace{f(x)}_{u}\underbrace{g(x)}_{v}-\int \underbrace{g(x)}_{v}\underbrace{f'(x)\dee{x}}_{\dee{u}}\\ \end{align*}1.7.2.2.
Solution.
1.7.2.3.
Solution.
Antidifferentiating both sides gives us:
\begin{align*} \int\diff{}{x}\left\{\dfrac{f(x)}{g(x)}\right\} \dee{x}&= \int\frac{g(x)f'(x)-f(x)g'(x)}{g^2(x)}\dee{x}\\ \dfrac{f(x)}{g(x)} +C&= \int\frac{f'(x)}{g(x)}\dee{x}-\int\frac{f(x)g'(x)}{g^2(x)}\dee{x}\\ \int\frac{f'(x)}{g(x)}\dee{x} &= \dfrac{f(x)}{g(x)}+\int\frac{f(x)g'(x)}{g^2(x)}\dee{x} \end{align*}1.7.2.4.
Solution.
1.7.2.5.
Solution.
1.7.2.6. (✳).
Solution.
1.7.2.7. (✳).
Solution.
1.7.2.8. (✳).
Solution.
1.7.2.9. (✳).
Solution.
1.7.2.10.
Solution.
Now, we take \(u=x^2\) and \(\dee{v}=e^x\dee{x}\text{,}\) so \(\dee{u}=2x\dee{x}\) and \(v=e^x\text{.}\) We’re only using integration by parts on the actual integral--the rest of the function stays the way it is.
\begin{align*} &=x^3e^x -3 \left[\underbrace{x^2e^x}_{uv} - \int \underbrace{e^x}_{v} \cdot \underbrace{2x\dee{x}}_{\dee{u}}\right]\\ &=x^3e^x - 3x^2e^x + 6\int x e^x\dee{x}\\ \end{align*}Continuing, we take \(u=x\) and \(\dee{v}=e^x\dee{x}\text{,}\) so \(\dee{u}=\dee{x}\) and \(v=e^x\text{.}\) This is the step where the polynomial part of the integrand finally disappears.
\begin{align*} &=x^3e^x - 3x^2e^x + 6\left[\underbrace{xe^x}_{uv} - \int \underbrace{ e^x}_v \underbrace{\dee{x}}_{\dee{u}}\right]\\ &=x^3e^x-3x^2e^x+6xe^x - 6e^x+C\\ &=e^x\left(x^3-3x^2+6x - 6\right)+C \end{align*}1.7.2.11.
Solution.
Continuing our quest to differentiate away the logarithm, we use integration by parts with \(u=\log^2x\text{,}\) \(\dee{v} = x\dee{x}\text{,}\) \(\dee{u} = \dfrac{2}{x}\log x\dee{x}\text{,}\) and \(v = \dfrac{1}{2}x^2\text{.}\)
\begin{align*} &=\frac{1}{2}x^2\log^3x - \frac{3}{2}\left[ \underbrace{\frac{1}{2}x^2\log^2 x}_{uv} - \int \underbrace{x\log x \dee{x}}_{v\dee{u}} \right]\\ &=\frac{1}{2}x^2\log^3x - \frac{3}{4}x^2\log^2 x + \frac{3}{2}\int x\log x \dee{x}\\ \end{align*}One last integration by parts: \(u=\log x\text{,}\) \(\dee{v}=x\dee{x}\text{,}\) \(\dee{u}=\dfrac{1}{x}\dee{x}\text{,}\) and \(v = \dfrac{1}{2}x^2\text{.}\)
\begin{align*} &=\frac{1}{2}x^2\log^3x - \frac{3}{4}x^2\log^2 x + \frac{3}{2}\left[ \underbrace{\frac{1}{2}x^2\log x}_{uv} - \int \underbrace{\frac{1}{2}x\dee{x}}_{v\dee{u}} \right]\\ &=\frac{1}{2}x^2\log^3x - \frac{3}{4}x^2\log^2 x + \frac{3}{4}x^2\log x - \frac{3}{4}\int x\dee{x}\\ &=\frac{1}{2}x^2\log^3x - \frac{3}{4}x^2\log^2 x + \frac{3}{4}x^2\log x - \frac{3}{8}x^2+C \end{align*}1.7.2.12.
Solution.
Using integration by parts again, we want to be differentiating (not antidifferentiating) \(x\text{,}\) so we choose \(u=x\text{,}\) \(\dee{v}=\cos x \dee{x}\text{,}\) and then \(\dee{u}=\dee{x}\) (\(x\) went away!), \(v=\sin x\text{.}\)
\begin{align*} &=-x^2\cos x + 2\left[\underbrace{x\sin x}_{uv}-\int \underbrace{\sin x \dee{x}}_{v\dee{u}} \right]\\ &=-x^2\cos x + 2x\sin x+2\cos x +C\\ &=(2-x^2)\cos x + 2x\sin x +C \end{align*}1.7.2.13.
Solution.
1.7.2.14.
Solution.
Now we have nearly the situation of Question 10. We can repeatedly use integration by parts with \(u\) as the power of \(w\) to get rid of the polynomial part. We’ll start with \(u=w^2\) \(\dee{v}=e^w\dee{w}\text{,}\) \(\dee{u}=2w\dee{w}\text{,}\) and \(v=e^w\text{.}\)
\begin{align*} &=2\left[\underbrace{w^2e^w}_{uv} - \int \underbrace{2we^w\dee{w}}_{v\dee{u}} \right]\\ &=2w^2e^w - 4\int we^w \dee{w}\\ \end{align*}We use integration by parts again, this time with \(u=w\text{,}\) \(\dee{v}=e^w\dee{w}\text{,}\) \(\dee{u}=\dee{w}\text{,}\) and \(v=e^w\text{.}\)
\begin{align*} &=2w^2e^w - 4\left[\underbrace{we^w}_{uv} - \int\underbrace{ e^w\dee{w}}_{v\dee{u}}\right]\\ &=2w^2e^w - 4we^w +4e^w+C\\ &=e^w\left(2w^2 - 4w +4\right)+C\\ &=e^{\sqrt{s}}\left(2s - 4\sqrt{s} +4\right)+C \end{align*}1.7.2.15.
Solution.
- Solution 1: Following Example 1.7.8, we choose \(u=\log^2 x\) and \(\dee{v}=\dee{x}\text{,}\) so that \(\dee{u}=\frac{2}{x}\log x \dee{x}\) and \(v=x\text{.}\)\begin{align*} \int \log^2 x \dee{x}&= \underbrace{x\log^2x}_{uv} - \int \underbrace{ 2\log x \dee{x}}_{v\dee{u}}\\ \end{align*}
Here we can either use the antiderivative of logarithm from memory, or re-derive it. We do the latter, using integration by parts with \(u=\log x\text{,}\) \(\dee{v}=2\dee{x}\text{,}\) \(\dee{u}=\frac{1}{x}\dee{x}\text{,}\) and \(v=2x\text{.}\)
\begin{align*} &=x\log^2 x - \left[\underbrace{2x\log x}_{uv} - \int \underbrace{2\dee{x}}_{v\dee{u}}\right]\\ &=x\log^2 x - 2x\log x +2x+C \end{align*} - Solution 2: Our integrand is two functions multiplied together: \(\log x\) and \(\log x\text{.}\) So, we will use integration by parts with \(u=\log x\text{,}\) \(\dee{v}=\log x\text{,}\) \(\dee{u}=\frac{1}{x}\dee{x}\text{,}\) and (using the antiderivative of logarithm, found in Example 1.7.8 in the text) \(v=x\log x -x\text{.}\)\begin{align*} \int \log^2 x \dee{x}&=(\underbrace{\log x}_u) (\underbrace{x\log x - x}_{v}) -\int (\underbrace{x\log x-x}_v)\underbrace{ \frac{1}{x}\dee{x}}_{\dee{u}}\\ &= x\log^2 x - x\log x - \int \left(\log x -1\right)\dee{x}\\ &= x\log^2 x - x\log x - \left[(x\log x - x )-x\right]+C\\ &=x\log^2 x -2x\log x +2x+C \end{align*}
1.7.2.16.
Solution.
1.7.2.17. (✳).
Solution.
Using the substitution \(u=1-y^2\text{,}\) \(\dee{u}=-2y\dee{y}\text{,}\)
\begin{align*} &=y\arccos y -\frac{1}{2} \int u^{-1/2}\dee{u}\\ &=y\arccos y - u^{1/2} +C\\ &= y \arccos y - \sqrt{1-y^2} + C \end{align*}1.7.2.18. (✳).
Solution.
1.7.2.19.
Solution.
- Option 1: Option 1 seems likelier. Let’s see how it plays out. We use integration by parts with \(u=\arctan x\text{,}\) \(\dee{v}=x^2 \dee{x}\text{,}\) \(\dee{u}=\frac{\dee{x}}{1+x^2}\text{,}\) and \(v=\frac{1}{3}x^3\text{.}\)\begin{align*} \displaystyle\int x^2\arctan x \dee{x}&=\underbrace{\frac{x^3}{3}\arctan x}_{uv} - \int \underbrace{\frac{x^3}{3(1+x^2)}\dee{x}}_{v\dee{u}}\\ &=\frac{x^3}{3}\arctan x- \frac{1}{3}\int \frac{x^3}{1+x^2}\dee{x}\\ \end{align*}
This is starting to look like a candidate for a substitution! Let’s try the denominator, \(s =1+x^2\text{.}\) Then \(\dee{s} = 2x\dee{x}\text{,}\) and \(x^2 = s-1\text{.}\)
\begin{align*} &= \frac{x^3}{3}\arctan x- \frac{1}{6}\int \frac{x^2}{\textcolor{red}{1+x^2}}\cdot \textcolor{blue}{2x\dee{x}}\\ &=\frac{x^3}{3}\arctan x- \frac{1}{6}\int \frac{s-1}{\textcolor{red}{s}}\textcolor{blue}{\dee{s}}\\ &=\frac{x^3}{3}\arctan x- \frac{1}{6} \int 1 - \frac{1}{s}\dee{s}\\ &=\frac{x^3}{3}\arctan x- \frac{1}{6}s + \frac{1}{6}\log|s|+C\\ &=\frac{x^3}{3}\arctan x- \frac{1}{6}(1+x^2) + \frac{1}{6}\log(1+x^2)+C \end{align*} - Option 2: What if we had tried the other option? That is, \(u=x^2\text{,}\) \(\dee{u} = 2x\dee{x}\text{,}\) \(\dee{v} = \arctan x\text{,}\) and \(v = x\arctan x - \frac{1}{2}\log(1+x^2)\text{.}\) It’s not always the case that both options work, but sometimes they do. (They are almost never of equal difficulty.) This solution takes advantage of two previously hard-won results: the antiderivatives of logarithm and arctangent.
Adding \(\quad 2\displaystyle\int x^2\arctan x \dee{x} \quad \) to both sides:
\begin{align*} 3\int x^2\arctan x \dee{x}&=x^3\arctan x - \frac{x^2}{2}\log(1+x^2) \\ &\hskip0.5in+ \int x\log(1+x^2)\dee{x}\\ \int x^2\arctan x \dee{x}&=\frac{x^3}{3}\arctan x - \frac{x^2}{6}\log(1+x^2) \\ &\hskip0.5in+ \frac{1}{3}\int x\log(1+x^2)\dee{x}\\ \end{align*}Using the substitution \(s=1+x^2\text{,}\) \(\dee{s}=2x\dee{x}\text{:}\)
\begin{align*} &=\frac{x^3}{3}\arctan x - \frac{x^2}{6}\log(1+x^2) + \frac{1}{6}\int\log s\dee{s}\\ \end{align*}Using the antiderivative of logarithm found in Example 1.7.8,
\begin{align*} &=\frac{x^3}{3}\arctan x - \frac{x^2}{6}\log(1+x^2) + \frac{1}{6}\left(s\log s - s\right)+C\\ &=\frac{x^3}{3}\arctan x - \frac{x^2}{6}\log(1+x^2) + \frac{1}{6}\big((1+x^2)\log (1+x^2) \\ &\hskip3in- (1+x^2)\big)+C\\ &=\frac{x^3}{3}\arctan x + \left[-\frac{x^2}{6}+\frac{1+x^2}{6}\right]\log(1+x^2)-\frac{1}{6}(1+x^2)+C\\ &=\frac{x^3}{3}\arctan x +\frac{1}{6}\log(1+x^2)-\frac{1}{6}(1+x^2)+C \end{align*}1.7.2.20.
Solution.
Similarly to our first integration by parts, we use \(u=\sin(2x)\text{,}\) \(\dee{v}=e^{x/2}\dee{x}\text{,}\) \(\dee{u}=2\cos(2x)\dee{x}\text{,}\) and \(v=2e^{x/2}\text{.}\)
\begin{align*} &=2e^{x/2}\cos(2x)+4\left[ \underbrace{2e^{x/2}\sin(2x)}_{uv} -\int\underbrace{4e^{x/2}\cos(2x)\dee{x}}_{v\dee{u}} \right] \end{align*}We add \(\quad 16\displaystyle\int e^{x/2}\cos(2x)\dee{x} \quad\) to both sides.
\begin{align*} \color{red}{17\int e^{x/2}\cos(2x)\dee{x}}&=2e^{x/2}\cos(2x)+8e^{x/2}\sin(2x)+C\\ \int e^{x/2}\cos(2x)\dee{x}&=\frac{2}{17}e^{x/2}\cos(2x)+\frac{8}{17}e^{x/2}\sin(2x)+C \end{align*}1.7.2.21.
Solution.
- Solution 1: This question looks like a substitution, since we have an “inside function.” So, let’s see where that leads: let \(u=\log x\text{.}\) Then \(\dee{u}=\dfrac{1}{x}\dee{x}\text{.}\) We don’t see this right away in our function, but we can bring it into the function by multiplying and dividing by \(x\text{,}\) and noting from our substitution that \(e^u=x\text{.}\)\begin{align*} \int \sin(\log x)\dee{x}&=\int \frac{x\sin(\log x)}{x}\dee{x}\\ &=\int e^u \sin u \dee{u}\\ \end{align*}
Using the result of Example 1.7.11:
\begin{align*} &=\frac{1}{2}e^u\left(\sin u - \cos u\right)+C\\ &=\frac{1}{2}e^{\log x}\left(\sin(\log x) - \cos (\log x)\right)+C\\ &=\frac{1}{2}x \left(\sin(\log x) - \cos (\log x)\right)+C \end{align*} -
Solution 2: It’s not clear how to antidifferentiate the integrand, but we can certainly differentiate it. So, keeping in mind the method of Example 1.7.11 in the text, we take \(u=\sin(\log x)\) and \(\dee{v}=\dee{x}\text{,}\) so \(\dee{u}=\frac{1}{x}\cos(\log x)\dee{x}\) and \(v=x\text{.}\)\begin{align*} \int \sin(\log x)\dee{x}&= \underbrace{x\sin(\log x)}_{uv} - \int \underbrace{ \cos(\log x)\dee{x} }_{v\dee{u}}\\ \end{align*}
Continuing on, we again use integration by parts, with \(u=\cos(\log x)\text{,}\) \(\dee{v}=\dee{x}\text{,}\) \(\dee{u}=-\frac{1}{x}\sin(\log x)\dee{x}\text{,}\) and \(v=x\text{.}\)
\begin{align*} &=x\sin(\log x) - \bigg[\underbrace{x\cos(\log x)}_{uv} + \int\underbrace{\sin(\log x)}_{-v\dee{u}}\dee{x} \bigg]\\ \end{align*}That is, we have
\begin{align*} \int \sin(\log x)\dee{x}&=x\left[\sin(\log x) - \cos(\log x)\right] - \int \sin(\log x)\dee{x}+C\\ \end{align*}Adding \(\quad \int \sin(\log x)\dee{x}\) to both sides,
\begin{align*} 2\int \sin(\log x)\dee{x}&=x\left[\sin(\log x) - \cos(\log x)\right]+C\\ \int \sin(\log x)\dee{x}&=\frac{x}{2}\left[\sin(\log x) - \cos(\log x)\right]+C \end{align*}Remark: remember that \(C\) is a stand-in for “we can add any real constant”. Since \(C\) can be any number in \((-\infty,\infty)\text{,}\) also \(\frac{C}{2}\) can be any number in \((-\infty,\infty)\text{.}\) So, rather than write \(\frac{C}{2}\) in the last line, we re-named \(\frac{C}{2}\) to \(C\text{.}\)
1.7.2.22.
Solution.
This is similar to the integral \(\displaystyle\int xe^x \dee{x}\text{,}\) which we saw in Example 1.7.1. Let’s write \(2=e^{\log 2}\) to take advantage of the easy integrability of \(e^x\text{.}\)
\begin{align*} &=\int x\cdot e^{x\log 2} \dee{x}\\ \end{align*}We use integration by parts with \(u=x\text{,}\) \(\dee{v}=e^{x\log 2}\dee{x}\text{;}\) \(\dee{u}=\dee{x}\text{,}\) \(v = \frac{1}{\log 2}e^{x\log 2}\text{.}\) (Remember \(\log 2\) is a constant. If you’d prefer, you can do a substitution with \(s= x\log 2\) first, to have a simpler exponent of \(e\text{.}\))
\begin{align*} &=\underbrace{\frac{x}{\log 2}e^{x\log 2}}_{uv} - \int \underbrace{\frac{1}{\log 2}e^{x\log 2}\dee{x}}_{v\dee{u}}\\ &=\frac{x}{\log 2}e^{x\log 2} - \frac{1}{(\log 2)^2}e^{x\log 2}+C\\ &=\frac{x}{\log 2}2^x - \frac{1}{(\log 2)^2}2^x+C \end{align*}1.7.2.23.
Solution.
Now we can use the substitution \(w=\cos x\text{,}\) \(\dee{w}=-\sin x \dee{x}\text{.}\)
\begin{align*} &=-2\int we^w\dee{w}\\ \end{align*}From here the integral should look more familiar. We can use integration by parts with \(u=w\text{,}\) \(\dee{v} = e^w\dee{w}\text{,}\) \(\dee{u}=\dee{w}\text{,}\) and \(v=e^w\text{.}\)
\begin{align*} &=-2\left[\underbrace{we^w}_{uv} - \int \underbrace{e^w\dee{w}}_{v\dee{u}}\right]\\ &=2e^w\left[1-w\right]+C\\ &=2e^{\cos x}[1-\cos x]+C \end{align*}1.7.2.24.
Solution.
1.7.2.25. (✳).
Solution.
Using the identity \(\sin^2 x + \cos^2 x =1 \text{,}\)
\begin{align*} &=-\sin^{n-1}x\ \cos x+(n-1)\int (1-\sin^2 x)\sin^{n-2}x\ \dee{x}\\ &=-\sin^{n-1}x\ \cos x+(n-1)\int\sin^{n-2}x\ \dee{x} -(n-1)\int\sin^{n}x\ \dee{x} \end{align*}1.7.2.26. (✳).
Solution.
- We cut \(R\) into thin horizontal strips of width \(\dee{y}\) as in the figure on the right above.
- When we rotate \(R\) about the \(y\)--axis, each strip sweeps out a thin washer
- whose inner radius is \(r_{in}=\tan y\) and outer radius is \(r_{out}=1\text{,}\) and
- whose thickness is \(\dee{y}\) and hence
- whose volume \(\pi(r_{out}^2 - r_{in}^2)\dee{y} = \pi(1-\tan^2 y)\dee{y}\text{.}\)
- As our bottommost strip is at \(y=0\) and our topmost strip is at \(y=\frac{\pi}{4}\) (since at the top \(x=1\) and \(x=\tan y\)), the total\begin{align*} \text{Volume} &= \int _0^{\frac{\pi}{4}} \pi(1-\tan^2 y)\dee{y} = \int _0^{\frac{\pi}{4}} \pi(2-\sec^2 y)\dee{y}\\ &=\pi\big[2y -\tan y\big]_0^{\frac{\pi}{4}}\\ &= \frac{\pi^2}{2}-\pi \end{align*}
1.7.2.27. (✳).
Solution.
1.7.2.28. (✳).
Solution.
1.7.2.29.
Solution.
- that our Riemann sum is a right Riemann sum (because we see \(i\text{,}\) not \(i-1\) or \(i-\frac{1}{2}\))
- \(\Delta x = \frac{2}{n}\) (because it is multiplied by the rest of the integrand, and also shows up multiplied by \(i\)),
- then \(x_i = a+i\Delta x = \frac{2}{n}i-1\text{,}\) which leads us to \(a=-1\) and
- \(f(x) = xe^x\text{.}\)
- Finally, since \(\Delta x = \frac{b-a}{n}=\frac{2}{n}\) and \(a=-1\text{,}\) we have \(b=1\text{.}\)
which we evaluate using integration by parts with \(u=x\text{,}\) \(\dee{v}=e^x\dee{x}\text{,}\) \(\dee{u}=\dee{x}\text{,}\) and \(v=e^x\text{.}\)
\begin{align*} &=\Big[\underbrace{xe^x}_{uv} \Big]_{-1}^1 - \int_{-1}^1 \underbrace{e^x\dee{x}}_{v\dee{u}}\\ &=\left(e+\frac{1}{e}\right) - \left(e-\frac{1}{e}\right) = \frac{2}{e} \end{align*}
1.8 Trigonometric Integrals
1.8.4 Exercises
1.8.4.1.
Solution.
If \(n=-1\text{,}\) then
\begin{align*} \int_0^{\pi/4} \sin x \cos^n x \dee{x}&= - \int_1^{1/\sqrt2} u^n \dee{u} = - \int_1^{1/\sqrt2} \frac{1}{u} \dee{u} \\ &=\bigg[ -\log|u|\bigg]_1^{1/\sqrt2}\\ & =-\log\left(\frac{1}{\sqrt2}\right) = \frac{1}{2}\log 2 \end{align*}1.8.4.2.
Solution.
Since \(n\) is positive, \(n-1 \neq -1\text{,}\) so we antidifferentiate using the power rule.
\begin{align*} &=\frac{u^n}{n}+C = \frac{1}{n}\sec^n x +C \end{align*}1.8.4.3.
Solution.
1.8.4.4. (✳).
Solution.
1.8.4.5. (✳).
Solution.
1.8.4.6. (✳).
Solution.
1.8.4.7.
Solution.
1.8.4.8.
Solution.
We can antidifferentiate the first integral right away. For the second integral, we use the identity \(\cos^2 \theta = \dfrac{1+\cos(2\theta)}{2}\text{,}\) with \(\theta=2x\text{.}\)
\begin{align*} &=\frac{1}{4}\Big[x - \sin(2x)\Big]_{0}^{\pi/3} + \frac{1}{8}\int_0^{\pi/3}(1+\cos(4x)) \dee{x}\\ &=\frac{1}{4}\left[\frac{\pi}{3}-\frac{\sqrt{3}}{2}\right]+\frac{1}{8}\Big[x+\frac{1}{4}\sin(4x)\Big]_0^{\pi/3}\\ &=\frac{1}{4}\left[\frac{\pi}{3}-\frac{\sqrt{3}}{2}\right]+\frac{1}{8}\left[\frac{\pi}{3}-\frac{\sqrt{3}}{8}\right]\\ &=\frac{\pi}{8} -\frac{9\sqrt3}{64} \end{align*}1.8.4.9.
Solution.
1.8.4.10.
Solution.
1.8.4.11.
Solution.
- Solution 1: Let’s use the substitution \(u=\tan x\text{,}\) \(\dee{u} = \sec^2 x \dee{x}\text{:}\)\begin{equation*} \int \tan x \sec^2 x \dee{x} = \int u \dee{u} = \frac{1}{2}u^2+C = \frac{1}{2}\tan^2 x +C \end{equation*}
- Solution 2: We can also use the substitution \(u=\sec x\text{,}\) \(\dee{u} = \sec x \tan x \dee{x}\text{:}\)\begin{equation*} \int\tan x \sec^2 x \dee{x} = \int u \dee{u} = \frac{1}{2}u^2+C = \frac{1}{2}\sec^2 x +C \end{equation*}
1.8.4.12. (✳).
Solution.
- Solution 1: Substituting \(u=\cos x\text{,}\) \(\dee{u}=-\sin x\,\dee{x}\text{,}\) \(\sin^2 x= 1-\cos^2x=1-u^2\text{,}\) gives\begin{align*} \int \tan^3 x \sec^5x \,\dee{x} &=\int\frac{\sin^3 x}{\cos^8 x}\,\dee{x} =\int\frac{(1-\cos^2 x)\sin x}{\cos^8 x}\,\dee{x}\\ &=-\int\frac{1-u^2}{u^8}\,\dee{u} =-\Big[\frac{u^{-7}}{-7}-\frac{u^{-5}}{-5}\Big]+C\\ &=\frac{1}{7}\sec^7 x -\frac{1}{5}\sec^5 x + C \end{align*}
- Solution 2: Alternatively, substituting \(u=\sec x\text{,}\) \(\dee{u}=\sec x\tan x\,\dee{x}\text{,}\) \(\tan^2 x= \sec^2x-1=u^2-1\text{,}\) gives\begin{align*} \int \tan^3 x \sec^5x \,\dee{x} &=\int \tan^2 x \sec^4x\ (\tan x\sec x)\,\dee{x}\\ &=\int (u^2-1) u^4\,\dee{u} =\Big[\frac{u^{7}}{7}-\frac{u^{5}}{5}\Big]+C\\ &=\frac{1}{7}\sec^7 x -\frac{1}{5}\sec^5 x + C \end{align*}
1.8.4.13. (✳).
Solution.
1.8.4.14.
Solution.
1.8.4.15.
Solution.
- Solution 1: As in Question 14, we have an odd power of tangent and at least one secant. So, as in strategy (2) of Section 1.8.2, we can use the substitution \(u=\sec x\text{,}\) \(\dee{u}=\sec x \tan x \dee{x}\text{,}\) and \(\tan^2 x = \sec^2 x -1=u^2-1\text{.}\)\begin{align*} \int \tan^3x\sec^2x \dee{x}&=\int \tan^2 x \sec x \cdot \sec x \tan x \dee{x}\\ &=\int(u^2-1)u \dee{u} = \int \left(u^3-u\right) \dee{u}\\ &=\frac{1}{4}u^4 - \frac{1}{2}u^2+C\\ &=\frac{1}{4}\sec^4 x - \frac{1}{2}\sec^2 x +C \end{align*}
- Solution 2: We have an even, strictly positve, power of \(\sec x\text{.}\) So, as in strategy (3) of Section 1.8.2, we can use the substitution \(u=\tan x\text{,}\) \(\dee{u}=\sec^2 x~\dee{x}\text{.}\)\begin{align*} \int \tan^3x\sec^2x~\dee{x}\amp=\int \tan^3 x \cdot \sec^2 x ~\dee{x}\\ \amp=\int u^3~\dee{u} \\ \amp=\frac{1}{4}u^4 + C\\ \amp=\frac{1}{4}\tan^4x +C \end{align*}
1.8.4.16.
Solution.
1.8.4.17.
Solution.
- Solution 1: Since the power of tangent is odd, let’s try to use the substitution \(u=\sec x\text{,}\) \(\dee{u} = \sec x \tan x \dee{x}\text{,}\) and \(\tan^2 x = \sec^2 x -1 = u^2-1\text{,}\) as in Questions 14 and 15. In order to make this work, we need to see \(\sec x \tan x \dee{x}\) in the integrand, so we do a little algebraic manipulation.\begin{align*} \int \tan^3 x \sec^{-0.7}x \dee{x}&= \int \dfrac{\tan^3 x}{\sec^{0.7 x}} \dee{x} = \int \dfrac{\tan^3 x}{\sec^{1.7 x}}\sec x \dee{x}\\ &=\int \frac{\tan^2x}{\sec^{1.7}x}\cdot \sec x \tan x \dee{x}\\ &=\int \frac{u^2-1}{u^{1.7}} \dee{u} = \int \left(u^{0.3}-u^{-1.7}\right) \dee{u}\\ &=\frac{u^{1.3}}{1.3} + \frac{1}{0.7u^{0.7}}+C\\ &=\frac{1}{1.3}\sec^{1.3}x + \frac{1}{0.7\sec^{0.7}x}+C\\ &=\frac{1}{1.3}\sec^{1.3}x + \frac{1}{0.7}\cos^{0.7}x+C \end{align*}
- Solution 2: Let’s convert the secants and tangents to sines and cosines.\begin{align*} \int \tan^3 x \sec^{-0.7}x \dee{x}&= \int \frac{\sin^3 x}{\cos^3 x}\cdot \cos^{0.7}x \dee{x}\\ &=\int\frac{\sin^3 x}{\cos^{2.3}x} \dee{x}=\int \frac{\sin^2 x}{\cos^{2.3}x}\cdot\sin x \dee{x}\\ \end{align*}
Using the substitution \(u=\cos x\text{,}\) \(\dee{u}=-\sin \dee{x}\text{,}\) and \(\sin^2 x = 1-\cos^2 x = 1-u^2\text{:}\)
\begin{align*} & = -\int\frac{1-u^2}{u^{2.3}} \dee{u} = \int \left(-u^{-2.3}+u^{-0.3}\right) \dee{u}\\ &=\frac{1}{1.3}u^{-1.3} + \frac{1}{0.7}u^{0.7}+C\\ &=\dfrac{1}{1.3}\sec^{1.3}x + \dfrac{1}{0.7}\cos^{0.7}x+C \end{align*}
1.8.4.18.
Solution.
Now we use the substitution \(u=\cos x\text{,}\) \(\dee{u}=-\sin x \dee{x}\text{,}\) and \(\sin^2 x = 1-\cos^2 x = 1-u^2\text{.}\)
\begin{align*} &=-\int\frac{(1-u^2)^2}{u^5} \dee{u} = \int \left(-u^{-5}+2u^{-3}-u^{-1}\right) \dee{u}\\ &=\frac{1}{4}u^{-4} - u^{-2}-\log|u|+C\\ &=\dfrac{1}{4}\sec^4 x - \sec^2 x - \log|\cos x|+C\\ &=\dfrac{1}{4}\sec^4 x - \sec^2 x + \log|\sec x|+C \end{align*}1.8.4.19.
Solution.
Note \(\tan(0)=0\text{,}\) and \(\tan(\pi/6)=1/\sqrt{3}\text{.}\)
\begin{align*} &=\int_0^{1/\sqrt{3}}(u^4-u^2+1)\dee{u} - \big[ x\big]_0^{\pi/6}\\ &=\left[\frac{1}{5}u^5 - \frac{1}{3}u^3+u\right]_0^{1/\sqrt{3}} - \frac{\pi}{6}\\ &=\frac{1}{5\sqrt{3}^5} - \frac{1}{3\sqrt{3}^3}+\frac{1}{\sqrt{3}}-\frac{\pi}{6}\\ &=\dfrac{41}{45\sqrt{3}} - \dfrac{\pi}{6} \end{align*}1.8.4.20.
Solution.
1.8.4.21.
Solution.
- Solution 1: Let’s use the substitution \(u=\sec x\text{,}\) \(\dee{u}=\sec x \tan x\dee{x}\text{.}\) In order to make this work, we need to see \(\sec x \tan x\) in the integrand, so we start with some algebraic manipulation.\begin{align*} \int \tan x \sqrt{\sec x}\left(\frac{\sqrt{\sec x}}{\sqrt{\sec x}}\right) \dee{x}&=\int \frac{1}{\sqrt{\sec x}}\sec x\tan x\dee{x}\\ &=\int \frac{1}{\sqrt{u}}\dee{u}=2\sqrt{u}+C\\ &=2\sqrt{\sec x}+C \end{align*}
- Solution 2: Let’s turn our secants and tangents into sines and cosines.\begin{align*} \int \tan x \sqrt{\sec x}\dee{x}&=\int \frac{\sin x}{\cos x\cdot\sqrt{\cos x}}\dee{x}=\int \frac{\sin x}{\cos^{1.5}x}\dee{x}\\ \end{align*}
We use the substitution \(u=\cos x\text{,}\) \(\dee{u}=-\sin x\dee{x}\text{.}\)
\begin{align*} &=\int -u^{-1.5}\dee{u}=\frac{2}{\sqrt{u}}+C\\ &=2\sqrt{\sec x}+C \end{align*}
1.8.4.22.
Solution.
1.8.4.23. (✳).
Solution.
1.8.4.24.
Solution.
Substitute \(u=\cos x\text{,}\) so \(\dee{u}=-\sin x\dee{x}\) and \(\sin^2 x = 1-\cos^2 x = 1-u^2\text{.}\)
\begin{align*} &=\int \frac{\sin^4 x}{\cos^3 x} \sin x\dee{x} =-\int \frac{(1-u^2)^2}{u^3}\dee{u}\\ &=-\int \frac{1-2u^2+u^4}{u^3}\dee{u}= \int \left(-\frac{1}{u^3}+\frac{2}{u}-u\right)\dee{u}\\ &=\frac{1}{2u^2}+2\log|u|-\frac{1}{2}u^2+C\\ &=\frac{1}{2\cos^2 x}+2\log|\cos x|-\frac{1}{2}\cos^2 x +C \end{align*}1.8.4.25.
Solution.
1.8.4.26.
Solution.
1.8.4.27.
Solution.
- Solution 1: We begin with the obvious substitution, \(w=e^x\text{,}\) \(\dee{w}=e^x \dee{w}\text{.}\)\begin{align*} \int e^x\sin(e^x)\cos(e^x) \dee{x}&= \int \sin w \cos w \dee{w}\\ \end{align*}
Now we see another substitution, \(u=\sin w\text{,}\) \(\dee{u}=\cos w\dee{w}\text{.}\)
\begin{align*} &=\int u\dee{u}=\frac{1}{2}u^2+C=\frac{1}{2}\sin^2 w +C\\ &=\frac{1}{2}\sin^2(e^x)+C \end{align*} - Solution 2: Notice that \(\diff{}{x}\{\sin(e^x)\} = e^x \cos(e^x)\text{.}\) This suggests to us the substitution \(u=\sin(e^x)\text{,}\) \(\dee{u} = e^x \cos(e^x)\dee{x}\text{.}\)\begin{align*} \int e^x\sin(e^x)\cos(e^x) \dee{x}&= \int u\dee{u} =\frac{1}{2}u^2+C = \frac{1}{2}\sin^2(e^x)+C \end{align*}
1.8.4.28.
Solution.
We use integration by parts with \(u=(1-s^2)\text{,}\) \(\dee{v}=\sin s \dee{s}\text{;}\) \(\dee{u}=-2s\dee{s}\text{,}\) and \(v = -\cos s\text{.}\)
\begin{align*} &=-\left[-(1-s^2)\cos s - \int 2s\cos s \dee{s}\right]\\ &= (1-s^2)\cos s + \int 2s\cos s \dee{s}\\ \end{align*}We integrate by parts again, with \(u=2s\text{,}\) \(\dee{v}=\cos s \dee{s}\text{;}\) \(\dee{u}=2\dee{s}\text{,}\) and \(v=\sin s\text{.}\)
\begin{align*} &=(1-s^2)\cos s + 2s\sin s - \int 2\sin s\dee{s}\\ &=(1-s^2)\cos s + 2s\sin s +2\cos s +C\\ &=\sin^2 x\cdot\cos (\cos x) + 2\cos x\cdot\sin (\cos x) +2\cos (\cos x) +C\\ &=(\sin^2x+2)\cos (\cos x) + 2\cos x\cdot\sin (\cos x) +C \end{align*}1.8.4.29.
Solution.
- Solution 1: Before we choose parts, let’s use the identity \(\sin(2x) = 2\sin x \cos x\text{.}\)\begin{align*} \int x\sin x \cos x \dee{x}&=\frac{1}{2}\int x \sin(2x)\dee{x}\\ \end{align*}
Now let \(u= x\text{,}\) \(\dee{v}=\sin(2x)\dee{x}\text{;}\) \(\dee{u}=\dee{x}\text{,}\) and \(v=-\frac{1}{2}\cos (2x) \text{.}\) Using integration by parts:
\begin{align*} &=\frac{1}{2}\left[-\frac{x}{2}\cos (2x) +\frac{1}{2} \int \cos (2x) \dee{x}\right]\\ &=-\frac{x}{4}\cos (2x) +\frac{1}{8}\sin (2x) +C\\ &=-\frac{x}{4}(1-2\sin^2x) +\frac{1}{4}\sin x\cos x +C\\ &=-\frac{x}{4} + \frac{x}{2}\sin^2x+\frac{1}{4}\sin x \cos x +C \end{align*} - Solution 2: If we let \(u=x\text{,}\) then \(\dee{u}=\dee{x}\text{,}\) and this seems desirable for integration by parts. If \(u=x\text{,}\) then \(\dee{v} = \sin x \cos x \dee{x}\text{.}\) To find \(v\) we can use the substitution \(u=\sin x\text{,}\) \(\dee{u}=\cos x \dee{x}\text{.}\)\begin{align*} v=\int \sin x \cos x \dee{x}&=\int u \dee{u} = \frac{1}{2}u^2+C = \frac{1}{2}\sin^2 x +C\\ \end{align*}
So, we take \(v = \frac{1}{2}\sin^2 x\text{.}\) Now we can apply integration by parts to our original integral.
\begin{align*} \int x\sin x \cos x \dee{x}&=\frac{x}{2}\sin^2 x - \int \frac{1}{2}\sin^2 x \dee{x}\\ \end{align*}Apply the identity \(\sin^2x = \dfrac{1-\cos(2x)}{2}\text{.}\)
\begin{align*} &=\frac{x}{2}\sin^2 x - \frac{1}{4}\int 1-\cos(2 x) \dee{x}\\ &=\frac{x}{2}\sin^2 x - \frac{x}{4} +\frac{1}{8}\sin(2 x)+C\\ &=\frac{x}{2}\sin^2 x - \frac{x}{4} +\frac{1}{4}\sin x \cos x+C \end{align*} - Solution 3: Let \(u=x\sin x\) and \(\dee{v}=\cos x \dee{x}\text{;}\) then \(\dee{u} = (x\cos x + \sin x)\dee{x}\) and \(v = \sin x\text{.}\)\begin{align*} \int x\sin x \cos x \dee{x}&=x\sin^2 x - \int \sin x(x\cos x + \sin x)\dee{x}\\ &=x\sin^2 x - \int x \sin x \cos x\dee{x} - \int \sin^2 x\dee{x}\\ \end{align*}Since \(C\) is an arbitrary constant that can take any number in \((-\infty,\infty)\text{,}\) also \(\frac{C}{2}\) is an arbitrary constant that can take any number in \((-\infty,\infty)\text{,}\) so we’re free to rename \(\frac{C}{2}\) to \(C\text{.}\)
Apply the identity \(\sin^2 x = \dfrac{1-\cos(2x)}{2}\) to the second integral.
\begin{align*} &=x\sin^2 x - \int x \sin x \cos x\dee{x} - \int \dfrac{1-\cos(2x)}{2}\dee{x}\\ &=x\sin^2 x - \int x \sin x \cos x\dee{x} - \frac{x}{2} +\frac{1}{4}\sin(2x)+C\\ \end{align*}So, we have the equation
\begin{align*} \color{red}{\int x\sin x \cos x \dee{x}}&=x\sin^2 x -\textcolor{red}{ \int x \sin x \cos x\dee{x}} \\ &\hskip1in- \frac{x}{2} + \frac{1}{4}\sin(2x)+C\\ \color{red}{2\int x\sin x \cos x \dee{x}} &=x\sin^2 x - \frac{x}{2} + \frac{1}{4}\sin(2x)+C\\ \int x\sin x \cos x \dee{x}&=\frac{x}{2}\sin^2 x - \frac{x}{4} + \frac{1}{8}\sin(2x)+\frac{C}{2}\\ &=\frac{x}{2}\sin^2 x - \frac{x}{4} + \frac{1}{4}\sin x\cos x+\frac{C}{2} \end{align*}
1.9 Trigonometric Substitution
1.9.2 Exercises
1.9.2.1. (✳).
Solution.
They have the following forms:
\begin{align*} &\mbox{constant } - \mbox{ function} &&\mbox{function } + \mbox{ constant} & &\mbox{function } - \mbox{ constant} \end{align*}What we want is a substitution that gives us
\begin{align*} 9x^2-16&=16\sec^2\theta - 16\\ \mbox{So,}\qquad 9x^2&=16\sec^2\theta\\ x &= \frac{4}{3}\sec\theta\\ \end{align*}Using this substitution,
\begin{align*} \sqrt{9x^2-16}&=\sqrt{16\sec^2\theta-16}\\ &=\sqrt{16\tan^2\theta}\\ &=4|\tan\theta| \end{align*}Using this substitution,
\begin{align*} \sqrt{1-4x^2}&=\sqrt{1-\sin^2\theta}\\ &=\sqrt{\cos^2\theta}\\ &=|\cos\theta| \end{align*}What we want is a substitution that gives us
\begin{align*} 25+x^2&=25\tan^2\theta+25\\ \mbox{So,}\qquad x^2&=25\tan^2\theta\\ x &= 5\tan\theta\\ \end{align*}Using this substitution,
\begin{align*} (25+x^2)^{-5/2}&=(25+25\tan^2\theta)^{-5/2}\\ &=(25\sec^2\theta)^{-5/2}\\ &=(5|\sec\theta|)^{-5} \end{align*}1.9.2.2.
Solution.
-
The quadratic function under the square root is \(x^2-4x+1\text{.}\) To complete the square, we match the non-constant terms to those of a perfect square.\begin{align*} (ax+b)^2&=a^2x^2+2abx+b^2\\ \textcolor{red}{x^2}-\textcolor{blue}{4x}+1&=\textcolor{red}{a^2x^2} + \textcolor{blue}{2abx} +b^2 + c \quad\mbox{for some constant } c \end{align*}
- Looking at the leading term tells us \(a=1\text{.}\)
- Then the second term tells us \(-4=2ab=2b\text{,}\) so \(b=-2\text{.}\)
- Finally, the constant terms give us \(1=b^2+c=4+c\text{,}\) so \(c=-3\text{.}\)
\begin{align*} \int \dfrac{1}{\sqrt{x^2-4x+1}}\dee{x} &=\int \dfrac{1}{\sqrt{(x-2)^2-3}}\dee{x}\\ &=\int \dfrac{1}{\sqrt{(x-2)^2-\sqrt{3}^2}}\dee{x} \end{align*}So we use the substitution \((x-2) = \sqrt{3}\sec u\text{,}\) which eliminates the square root:\begin{equation*} \sqrt{\left(x-2\right)^2-3}=\sqrt{3\sec^2 u - 3} = \sqrt{3\tan^2 u} = \sqrt{3}|\tan u| \end{equation*} -
The quadratic function under the square root is \(-x^2+2x+4=-[x^2-2x-4]\text{.}\) To complete the square, we match the non-constant terms to those of a perfect square. We factored out the negative to make things a little easier — don’t forget to put it back in before choosing a substitution!\begin{align*} (ax+b)^2&=a^2x^2+2abx+b^2\\ \textcolor{red}{x^2}-\textcolor{blue}{2x}-4&=\textcolor{red}{a^2x^2} + \textcolor{blue}{2abx} +b^2 + c \quad\mbox{for some constant } c \end{align*}
- Looking at the leading term tells us \(a=1\text{.}\)
- Then the second term tells us \(-2=2ab=2b\text{,}\) so \(b=-1\text{.}\)
- Finally, the constant terms give us \(-4=b^2+c=1+c\text{,}\) so \(c=-5\text{.}\)
- Then\begin{align*} -x^2+2x+4 &= -[x^2-2x-4]=-[(x-1)^2-5]\\ &=5-(x-1)^2 \end{align*}
\begin{align*} \int \dfrac{(x-1)^6}{(-x^2+2x+4)^{3/2}}\dee{x} &=\int \dfrac{(x-1)^6}{(5-(x-1)^2)^{3/2}}\dee{x}\\ &=\int \dfrac{(x-1)^6}{\left(\sqrt{5}^2-(x-1)^2\right)^{3/2}}\dee{x} \end{align*}So we use the substitution \((x-1) = \sqrt{5}\sin u\text{,}\) which eliminates the square root (fractional power):\begin{align*} {(5-(x-1)^2)}^{3/2}&={(5-5\sin^2u)}^{3/2} = {(5\cos^2 u)}^{3/2}\\ &= 5\sqrt{5}|\cos^3 u| \end{align*} - The quadratic function under the square root is \(4x^2+6x+10\text{.}\) To complete the square, we match the non-constant terms to those of a perfect square.\begin{align*} (ax+b)^2&=a^2x^2+2abx+b^2\\ \textcolor{red}{4x^2}+\textcolor{blue}{6x}+10&=\textcolor{red}{a^2x^2} + \textcolor{blue}{2abx} +b^2 + c \quad\mbox{for some constant $c$} \end{align*}
- Looking at the leading term tells us \(a=2\text{.}\)
- Then the second term tells us \(6=2ab=4b\text{,}\) so \(b=\frac{3}{2}\text{.}\)
- Finally, the constant terms give us \(10=b^2+c=\frac{9}{4}+c\text{,}\) so \(c=\frac{31}{4}\text{.}\)
\begin{align*} \int \dfrac{1}{\sqrt{4x^2+6x+10}}\dee{x} &=\int \dfrac{1}{\sqrt{\left(2x+\frac{3}{2}\right)^2+\frac{31}{4}}}\dee{x}\\ &=\int \dfrac{1}{\sqrt{\left(2x+\frac{3}{2}\right)^2+\left(\frac{\sqrt{31}}{2}\right)^2}}\dee{x} \end{align*}So we use the substitution \(\left(2x+\frac{3}{2}\right) =\frac{\sqrt{31}}{2}\tan u\text{,}\) which eliminates the square root:\begin{align*} \sqrt{\left(2x+\frac{3}{2}\right)^2+\frac{31}{4}} &=\sqrt{\frac{31}{4}\tan^2 u +\frac{31}{4}} = \sqrt{\frac{31}{4}\sec^2 u} \\ &=\frac{\sqrt{31}}{2}|\sec u| \end{align*} - The quadratic function under the square root is \(x^2-x\text{.}\) To complete the square, we match the non-constant terms to those of a perfect square.\begin{align*} (ax+b)^2&=a^2x^2+2abx+b^2\\ \textcolor{red}{x^2}-\textcolor{blue}{x}&=\textcolor{red}{a^2x^2} + \textcolor{blue}{2abx} +b^2 + c \quad\mbox{for some constant $c$} \end{align*}
- Looking at the leading term tells us \(a=1\text{.}\)
- Then the second term tells us \(-1=2ab=2b\text{,}\) so \(b=-\frac{1}{2}\text{.}\)
- Finally, the constant terms give us \(0=b^2+c=\frac{1}{4}+c\text{,}\) so \(c=-\frac{1}{4}\text{.}\)
\begin{align*} \int \sqrt{x^2-x}\dee{x} &=\int \sqrt{\left(x-\frac{1}{2}\right)^2 -\frac{1}{4}}\dee{x}\\ &=\int \sqrt{\left(x-\frac{1}{2}\right)^2 -\left(\frac{1}{2}\right)^2}\dee{x} \end{align*}So we use the substitution \((x-1/2) = \frac{1}{2}\sec u\text{,}\) which eliminates the square root:\begin{equation*} \sqrt{\left(x-\frac{1}{2}\right)^2-\frac{1}{4}}=\sqrt{\frac{1}{4}{\sec\vphantom{|}}^2 u - \frac{1}{4}} = \sqrt{\frac{1}{4}\tan^2 u} =\frac{1}{2}|\tan u| \end{equation*}
1.9.2.3.
Solution.
-
If \(\sin\theta=\dfrac{1}{20}\) and \(\theta\) is between 0 and \(\pi/2\text{,}\) then we can draw a right triangle with angle \(\theta\) that has opposite side length 1, and hypotenuse length 20. By the Pythagorean Theorem, the adjacent side has length \(\sqrt{20^2-1^2}=\sqrt{399}\text{.}\) So, \(\cos\theta = \dfrac{\mathrm{adj}}{\mathrm{hyp}}=\dfrac{\sqrt{399}}{20}\text{.}\)We can do a quick “reasonableness” check here: \(\frac{1}{20}\) is pretty close to 0, so we might expect \(\theta\) to be pretty close to 0, and so \(\cos \theta\) should be pretty close to 1. Indeed it is: \(\dfrac{\sqrt{399}}{20}\approx \dfrac{\sqrt{400}}{20}=\dfrac{20}{20}=1\text{.}\)Alternatively, we can solve this problem using identities.\begin{align*} \sin^2 \theta + \cos^2 \theta &=1\\ \left(\frac{1}{20}\right)^2+ \cos^2 \theta &=1\\ \cos\theta &= \pm\sqrt{1-\frac{1}{400}}=\pm\frac{\sqrt{399}}{20}\\ \end{align*}
Since \(0 \leq \theta \leq \frac{\pi}{2}\text{,}\) \(\cos\theta \geq 0\text{,}\) so
\begin{align*} \cos\theta &= \frac{\sqrt{399}}{20} \end{align*} -
If \(\tan\theta=7\) and \(\theta\) is between 0 and \(\pi/2\text{,}\) then we can draw a right triangle with angle \(\theta\) that has opposite side length 7 and adjacent side length 1. By the Pythagorean Theorem, the hypotenuse has length \(\sqrt{7^2+1^2} = \sqrt{50}=5\sqrt{2}\text{.}\) So, \(\csc\theta = \dfrac{\mathrm{hyp}}{\mathrm{opp}}=\dfrac{5\sqrt{2}}{7}\text{.}\)Again, we can do a quick reasonableness check. Since 7 is much larger than 1, the triangle we’re thinking of doesn’t look much like the triangle in our standardized picture above: it’s really quite tall, with a small base. So, the opposite side and hypotenuse are pretty close in length. Indeed, \(\dfrac{5\sqrt{2}}{7}\approx 7.071\text{,}\) so this dimension seems reasonable.
-
If \(\sec\theta=\dfrac{\sqrt{x-1}}{2}\) and \(\theta\) is between 0 and \(\pi/2\text{,}\) then we can draw a right triangle with angle \(\theta\) that has hypotenuse length \(\sqrt{x-1}\) and adjacent side length 2. By the Pythagorean Theorem, the opposite side has length \(\sqrt{\sqrt{x-1}^2 - 2^2} = \sqrt{x-1-4}=\sqrt{x-5}\text{.}\) So, \(\tan\theta = \dfrac{\mathrm{opp}}{\mathrm{adj}}=\dfrac{\sqrt{x-5}}{2}\text{.}\)We can also solve this using identities. Note that since \(\sec\theta\) exists, \(\theta \neq \frac{\pi}{2}\text{.}\)\begin{align*} \tan^2\theta+1&=\sec^2\theta\\ \tan^2\theta+1&=\left(\frac{\sqrt{x-1}}{2}\right)^2=\frac{x-1}{4}\\ \tan\theta &= \pm\sqrt{\frac{x-1}{4}-1} = \pm\frac{\sqrt{x-5}}{2}\\ \end{align*}
Since \(0 \leq \theta \lt \frac{\pi}{2}\text{,}\) \(\tan\theta \geq 0\text{,}\) so
\begin{align*} \tan\theta &= \frac{\sqrt{x-5}}{2} \end{align*}
1.9.2.4.
Solution.
-
Let \(\theta = \arccos \left(\frac{x}{2}\right)\text{.}\) That is, \(\cos(\theta) = \frac{x}{2}\text{,}\) and \(0 \leq \theta \leq \pi\text{.}\) Then we can draw the corresponding right triangle with angle \(\theta\) with adjacent side of signed length \(x\) (we note that if \(\theta \gt \frac{\pi}{2}\text{,}\) then \(x\) is negative) and hypotenuse of length \(2\text{.}\) By the Pythagorean Theorem, the opposite side of the triangle has length \(\sqrt{4-x^2}\text{.}\)So,\begin{equation*} \sin\left(\arccos \left(\frac{x}{2}\right)\right)=\sin \theta = \frac{\mathrm{opp}}{\mathrm{hyp}} = \frac{\sqrt{4-x^2}}{2} \end{equation*}
-
Let \(\theta = \arctan \left(\frac{1}{\sqrt{3}}\right)\text{.}\) That is, \(\tan(\theta) = \frac{1}{\sqrt{3}}\text{,}\) and \(-\frac{\pi}{2} \leq \theta \leq \frac{\pi}{2}\text{.}\)
- Solution 1: Then \(\theta = \dfrac{\pi}{6}\text{,}\) so \(\sin\theta = \dfrac{1}{2}\text{.}\)
-
Solution 2: Then we can draw the corresponding right triangle with angle \(\theta\) with opposite side of length \(1\) and adjacent side of length \(\sqrt{3}\text{.}\) By the Pythagorean Theorem, the hypotenuse of the triangle has length \(\sqrt{\sqrt{3}^2+1^2}=2\text{.}\)So,\begin{equation*} \sin\left(\arctan \left(\frac{1}{\sqrt{3}}\right)\right)=\sin \theta = \frac{\mathrm{opp}}{\mathrm{hyp}} = \frac{1}{2} \end{equation*}
-
Let \(\theta = \arcsin \left(\sqrt{x}\right)\text{.}\) That is, \(\sin(\theta) = \sqrt{x}\text{,}\) and \(-\frac{\pi}{2} \leq \theta \leq \frac{\pi}{2}\text{.}\) Then we can draw the corresponding right triangle with angle \(\theta\) with opposite side of length \(\sqrt{x}\) and hypotenuse of length \(1\text{.}\) By the Pythagorean Theorem, the adjacent side of the triangle has length \(\sqrt{1-x}\text{.}\)So,\begin{equation*} \sec\left(\arcsin \left(\sqrt{x}\right)\right)=\sec \theta = \frac{\mathrm{hyp}}{\mathrm{adj}} = \frac{1}{\sqrt{1-x}} \end{equation*}
1.9.2.5. (✳).
Solution.
1.9.2.6. (✳).
Solution.
-
Solution 1: As in Question 5, substitute \(x=2\tan u\text{,}\) \(\dee{x}=2 \sec^2u\,\dee{u}\text{.}\) Note that when \(x=4\) we have \(4=2\tan u\text{,}\) so that \(\tan u=2\text{.}\)\begin{align*} \int_0^4 \frac{1}{{(4+x^2)}^{3/2}}\,\dee{x} &=\int_0^{\arctan 2} \frac{1}{{(4+4\tan^2 u)}^{3/2}}\,2\sec^2 u\,\dee{u}\\ &=\int_0^{\arctan 2} \frac{2\sec^2 u}{{(2\sec u)}^{3}}\,\dee{u}\\ &=\frac{1}{4}\int_0^{\arctan 2} \frac{\sec^2u}{\sec^3u}\,\dee{u}\\ &=\frac{1}{4}\int_0^{\arctan 2} \cos u\,\dee{u}\\ &=\bigg[\frac{1}{4}\sin u \bigg]_0^{\arctan2}\\ &=\frac{1}{4} \big( \sin(\arctan 2) - 0 \big) = \frac{1}{2\sqrt{5}} \end{align*}To find \(\sin(\arctan 2)\text{,}\) we use the right triangle above, with angle \(u=\arctan 2\text{.}\) Since \(\tan u=2 = \dfrac{\mbox{opp}}{\mbox{adj}}\text{,}\) we label the opposite side as 2, and the adjacent side as 1. The Pythagorean Theorem tells us the hypotenuse has length \(\sqrt{5}\text{,}\) so \(\sin u = \dfrac{\mbox{opp}}{\mbox{hyp}} = \dfrac{2}{\sqrt{5}}\text{.}\)
- Solution 2: Using our result from Question 5,\begin{align*} \int_0^4 \frac{1}{{(4+x^2)}^{3/2}}\,\dee{x}&=\frac{1}{4}\left[ \dfrac x{\sqrt{x^2+4}}\right]_0^4\\ &=\frac{1}{4}\cdot \dfrac{4}{\sqrt{4^2+4}}=\frac{1}{2\sqrt{5}} \end{align*}
1.9.2.7. (✳).
Solution.
1.9.2.8. (✳).
Solution.
1.9.2.9.
Solution.
1.9.2.10. (✳).
Solution.
1.9.2.11. (✳).
Solution.
1.9.2.12. (✳).
Solution.
So,
\begin{align*} \int_0^{\pi/4}\cos^4\theta\dee{\theta} &=\int_0^{\pi/4}\Big(\frac{\cos4\theta}{8}+\frac{\cos2\theta}{2}+\frac{3}{8}\Big) \dee{\theta}\\ &=\left[\frac{\sin4\theta}{32}+\frac{\sin2\theta}{4}+\frac{3}{8}\theta\right]_0^{\pi/4}\\ &= \frac{1}{4}+\frac{3}{8}\cdot \frac{\pi}{4}\\ &=\frac{8+3\pi}{32} \end{align*}1.9.2.13.
Solution.
1.9.2.14. (✳).
Solution.
1.9.2.15. (✳).
Solution.
1.9.2.16.
Solution.
1.9.2.17. (✳).
Solution.
1.9.2.18.
Solution.
As \(x\gt 2\text{,}\) we have \(2x-3\gt 1\text{.}\) We use the substitution \(2x-3 = \sec \theta\) with \(0\le\theta \lt \frac{\pi}{2}\text{.}\) So \(2\,\dee{x}=\sec\theta\tan\theta~\dee{\theta}\) and \(\sqrt{(2x-3)^2-1}=\sqrt{\sec^2\theta-1}=\sqrt{\tan^2\theta}=\tan\theta\text{.}\)
\begin{align*} &=\frac{1}{2}\int\frac{1}{\sec^3\theta\sqrt{\sec^2\theta-1}}\sec\theta\tan\theta\dee{\theta}\\ &=\frac{1}{2}\int\frac{1}{\sec^3\theta\tan\theta}\sec\theta\tan\theta\dee{\theta}\\ &=\frac{1}{2}\int\frac{1}{\sec^2\theta}\dee{\theta}\\ &=\frac{1}{2}\int{\cos^2\theta}\dee{\theta}\\ &=\frac{1}{4}\int{\left(1+\cos(2\theta)\right)}\dee{\theta}\\ &=\frac14\left(\theta + \frac{1}{2}\sin(2\theta)\right)+C\\ &=\frac14\left(\theta + \sin\theta\cos\theta\right)+C\\ &=\frac{1}{4}\left(\arccos\left(\frac{1}{2x-3}\right) + \frac{\sqrt{4x^2-12x+8}}{(2x-3)^2}\right)+C \end{align*}1.9.2.19.
Solution.
1.9.2.20.
Solution.
1.9.2.21.
Solution.
We use the substitution \(x-1=\tan\theta\text{,}\) which implies \(\dee{x}=\sec^2\theta\dee{\theta}\) and \(x=\tan\theta+1\)
\begin{align*} &=\int \dfrac{(\tan\theta+1)^2}{\sqrt{(\tan\theta)^2+1}} \sec^2\theta\dee{\theta}\\ &=\int\frac{\textcolor{red}{\tan^2\theta}+2\tan\theta+\textcolor{red}1}{\sec\theta}\sec^2\theta\dee{\theta}\\ &=\int(\textcolor{red}{\sec^2\theta}+2\tan\theta)\sec\theta\dee{\theta}\\ &=\int\big( \sec^3\theta+2\tan\theta\sec\theta\big)\dee{\theta}\\ &=\frac{1}{2}\sec\theta\tan\theta + \frac{1}{2}\log|\sec\theta+\tan\theta|+2\sec\theta+C\\ &=\frac{1}{2}\sqrt{x^2-2x+2}(x-1) + \frac{1}{2}\log\left|\sqrt{x^2-2x+2}+x-1\right|\\ &\hskip2in+2\sqrt{x^2-2x+2}+C\\ &=\frac{3+x}{2}\sqrt{x^2-2x+2}+ \frac{1}{2}\log\left|\sqrt{x^2-2x+2}+x-1\right|+C \end{align*}1.9.2.22.
Solution.
- Looking at the leading term tells us \(a=\sqrt{3}\text{.}\)
- Then the second term tells us \(5=2ab=2\sqrt{3}b\text{,}\) so \(b=\frac{5}{2\sqrt3}\text{.}\)
- Finally, the constant terms give us \(0=b^2+c=\frac{25}{12}+c\text{,}\) so \(c=-\frac{25}{12}\text{.}\)
We use the substitution \(\sqrt{3}x + \frac{5}{2\sqrt3}=\frac{5}{2\sqrt{3}}\sec\theta\text{,}\) which leads to \(\sqrt{3}\dee{x} = \frac{5}{2\sqrt{3}}\sec\theta\tan\theta\dee{\theta}\text{,}\) i.e. \(\dee{x} = \frac{5}{6}\sec\theta\tan\theta\dee{\theta}\text{.}\)
\begin{align*} &=\int\frac{1}{\sqrt{\left(\frac{5}{2\sqrt3}\sec\theta\right)^2-\frac{25}{12}}}\cdot\frac{5}{6}\sec\theta\tan\theta\dee{\theta}\\ &=\int\frac{1}{\sqrt{\frac{25}{12}\sec^2\theta-\frac{25}{12}}}\cdot\frac{5}{6}\sec\theta\tan\theta\dee{\theta}\\ &=\int\frac{1}{\sqrt{\frac{25}{12}\tan^2\theta}}\cdot\frac{5}{6}\sec\theta\tan\theta\dee{\theta}\\ &=\int\frac{1}{{\frac{5}{2\sqrt{3}}\tan\theta}}\cdot\frac{5}{6}\sec\theta\tan\theta\dee{\theta}\\ &=\frac{1}{\sqrt3}\int\sec\theta\dee{\theta}\\ &=\frac{1}{\sqrt3}\log\left|\sec\theta+\tan\theta\right|+C\\ &=\frac{1}{\sqrt3}\log\left| \left(\frac{6}{5}x+1\right)+\frac{2}{5}\sqrt{9x^2+15x} \right|+C \end{align*}1.9.2.23.
Solution.
1.9.2.24.
Solution.
1.9.2.25.
Solution.
We use the substitution \(u=1-x^2\text{,}\) \(\dee{u}=-2x\dee{x}\text{.}\)
\begin{align*} &=-\int_{1}^{3/4} \frac{1}{u^{1/4}}\dee{u}\\ &=-\left[\frac{4}{3}u^{3/4}\right]_1^{3/4} = -\frac{4}{3}\left(\left(\frac{3}{4}\right)^{3/4}-1\right)\\ &=\frac{4}{3} - \sqrt[4]{\frac{4}{3}} \end{align*}We use the substitution \(x=\sin \theta\text{,}\) \(\dee{x} = \cos\theta\dee{\theta}\text{,}\) so \(\sqrt{1-x^2} = \sqrt{1-\sin^2\theta}=\cos \theta\text{.}\) Note \(\sin 0 =0\) and \(\sin\frac{\pi}{6}=\frac{1}{2}.\)
\begin{align*} &=2\pi\int_{0}^{\pi/6} \frac{\sin^2 \theta}{\cos \theta}\cos \theta\dee{\theta}\\ &=2\pi\int_{0}^{\pi/6} \sin^2 \theta\dee{\theta}\\ &=\pi\int_{0}^{\pi/6}\big(1- \cos(2 \theta)\big)\dee{\theta}\\ &=\pi\left[\theta - \frac{1}{2}\sin(2\theta)\right]_0^{\pi/6}\\ &=\pi\left(\frac{\pi}{6} - \frac{1}{2}\cdot \frac{\sqrt{3}}{2}\right)\\ &=\frac{\pi^2}{6} - \frac{\sqrt{3}\pi}{4} \end{align*}1.9.2.26.
Solution.
1.9.2.27.
Solution.
- We can save ourselves some trouble by applying logarithm rules before we differentiate.\begin{align*} &\log\left| \dfrac{1+x}{\sqrt{1-x^2}}\right|=\log|1+x| - \log|\sqrt{1-x^2}|\\ &\hskip0.5in=\log|1+x| - \frac{1}{2}\log|1-x^2|\\ &\hskip0.5in=\log|1+x| - \frac{1}{2}\log|(1+x)(1-x)|\\ &\hskip0.5in=\log|1+x| - \frac{1}{2}\log|1+x|- \frac{1}{2}\log|1-x| \end{align*}\begin{align*} &\diff{}{x}\left\{\log\left| \dfrac{1+x}{\sqrt{1-x^2}}\right|\right\}\\ &\hskip0.5in=\diff{}{x}\left\{\log|1+x| - \frac{1}{2}\log|1+x|- \frac{1}{2}\log|1-x|\right\}\\ &\hskip0.5in=\frac{1}{1+x} - \frac{1/2}{1+x}+ \frac{1/2}{1-x}\\ &\hskip0.5in= \frac{1/2}{1+x}+ \frac{1/2}{1-x}\\ &\hskip0.5in=\frac{1}{1-x^2} \end{align*}Notice this is the integrand from our work in blue.
- False: \(\displaystyle\int_{2}^{3} \frac{1}{1-x^2}\dee{x}\) is a number, because it is the area under a finite portion of a continuous curve. (We note that the integrand is continuous over the interval \([2,3]\text{,}\) although it is not continuous everywhere.) However, \(\left[\log\left| \dfrac{1+x}{\sqrt{1-x^2}}\right|\right]_{x=2}^{x=3}\) is not defined, since the denominator takes the square root of a negative number. So, these two expressions are not the same.
-
The work in the question is not correct. The most salient problem is that when we make the substitution \(x=\sin\theta\text{,}\) we restrict the possible values of \(x\) to \([-1,1]\text{,}\) since this is the range of the sine function. However, the original integral had no such restriction.How can we be sure we avoid this problem in the future? In the introductory text to Section 1.9 (before Example 1.9.1), the notes tell us that we are allowed to write our old variable as a function of a new variable (say \(x=s(u)\)) as long as that function is invertible to recover our original variable \(x\text{.}\) There is one very obvious reason why invertibility is necessary: after we antidifferentiate using our new variable \(u\text{,}\) we need to get it back in terms of our original variable, so we need to be able to recover \(x\text{.}\) Moreover, invertibility reconciles potential problems with domains: if an inverse function \(u=s^{-1}(x)\) exists, then for any \(x\text{,}\) there exists a \(u\) with \(s(u)=x\text{.}\) (This was not the case in the work for the question, because we chose \(x=\sin \theta\text{,}\) but if \(x=2\text{,}\) there is no corresponding \(\theta\text{.}\) Note, however, that \(x=\sin\theta\) is invertible over \([-1,1]\text{,}\) so the work is correct if we restrict \(x\) to those values.)Remark: in the next section, you will learn to use partial fractions to find \(\displaystyle\int \dfrac{1}{1-x^2}\dee{x} = \log|1+x|-\dfrac{1}{2}\log|1-x|\text{.}\) When \(-1 \lt x \lt 1\text{,}\) this is equivalent to \(\log\left| \dfrac{1+x}{\sqrt{1-x^2}}\right|\text{.}\)
1.9.2.28.
Solution.
1.10 Partial Fractions
1.10.4 Exercises
1.10.4.1.
Solution.
1.10.4.2. (✳).
Solution.
1.10.4.3. (✳).
Solution.
1.10.4.4.
Solution.
-
We start by dividing. The leading term of the numerator is \(x\) times the leading term of the denominator. The remainder is \(x+2\text{.}\)That is, \(x^3+2x+2 = x(x^2+1)+(x+2)\text{.}\) So,\begin{equation*} \frac{x^3+2x+2}{x^2+1} = x+\frac{x+2}{x^2+1} \end{equation*}
-
We start by dividing. The leading term of the numerator is \(3x^2\) times the leading term of the denominator.Then \(5x^2\) goes into \(10x^2\) twice, soOur remainder is 4. That is,\begin{equation*} \dfrac{15x^4+6x^3+34x^2+4x+20}{5x^2+2x+8} = 3x^2+2+\frac{4}{5x^2+2x+8}. \end{equation*}
-
We start by dividing. The leading term of the numerator is \(x^3\) times the leading term of the denominator.Then \(2x^2(2x)\) gives us \(4x^3\text{.}\)Finally, \(2x^2\) goes into \(12x^2\) six times.Since there is no remainder,\begin{equation*} \dfrac{2x^5+9x^3+12x^2+10x+30}{2x^2+5}=x^3+2x+6 \end{equation*}Remark: if we wanted to be pedantic about the question statement, we could write our final answer as \(x^3+2x+6+\frac{0}{x}\text{,}\) so that we are indeed adding a polynomial to a rational function whose numerator has degree strictly smaller than its denominator.
1.10.4.5.
Solution.
- The polynomial \(5x^3-3x^2-10x+6\) has a repeated pattern: the ratio of the first two coefficients is the same as the ratio of the last two coefficients. We can use this to factor.\begin{align*} 5x^3-3x^2-10x+6&=x^2(5x-3)-2(5x-3) \\ &= (x^2-2)(5x-3)\\ &=(x+\sqrt{2})(x-\sqrt{2})(5x-3) \end{align*}
- The polynomial \(x^4-3x^2-5\) has only even powers of \(x\text{,}\) so we can (temporarily) replace them with \(x^2=y\) to turn our quartic polynomial into a quadratic.\begin{align*} x^4-3x^2-5&=y^2-3y-5\\ \end{align*}
There’s no obvious factoring here, but we can find its roots, if any, using the quadratic equation.
\begin{align*} y&=\frac{3\pm\sqrt{3^2-4(1)(-5)}}{2}\\ &=\frac{3\pm\sqrt{29}}{2}\\ \mbox{So,}\ y^2-3y-5&=\left(y-\frac{3+\sqrt{29}}{2}\right)\left(y-\frac{3-\sqrt{29}}{2}\right)\\ \mbox{Therefore,}\ x^4-3x^2-5&=\left(x^2-\frac{3+\sqrt{29}}{2}\right)\left(x^2-\frac{3-\sqrt{29}}{2}\right)\\ \end{align*}We’d like to use the difference of two squares to factor these quadratic expressions. For this to work, the constants must be positive (so their square roots are real). Since \(\sqrt{29} \gt 3\text{,}\) only the first quadratic is factorable. The other is irreducible — it’s always positive, so it had no roots.
\begin{align*} x^4-3x^2-5&=\left(x+\sqrt{\frac{3+\sqrt{29}}{2}}\right)\left(x-\sqrt{\frac{3+\sqrt{29}}{2}}\right)\\ &\hskip2in\left(x^2+\frac{\sqrt{29}-3}{2}\right) \end{align*} -
Without seeing any obvious patterns, we start hunting for roots. Since we have all integer coefficients, if there are any integer roots, they will divide our constant term, \(-6\text{.}\) So, our candidates for roots are \(\pm1,\,\pm2,\,\pm3,\) and \(\pm6\text{.}\) To save time, we don’t need to know exactly the value of our polynomial at these points: only whether or not it is 0. Write \(f(x) = x^4 - 4x^3 - 10x^2 - 11x - 6\text{.}\)\begin{align*} \color{red}{f(-1)}&\color{red}{=0}& f(-2)&\neq 0& f(-3)&\neq 0& f(-6)&\neq 0\\ f(1)&\neq 0& f(2)&\neq 0& f(3)&\neq 0& \color{red}{f(6)}&\color{red}{= 0} \end{align*}Since \(x=-1\) and \(x=6\) are roots of our polynomial, it has factors \((x+1)\) and \((x-6)\text{.}\) Note \((x+1)(x-6) = x^2-5x-6\text{.}\) We use long division to figure out what else is lurking in our polynomial.So, \(x^4-4x^3-10x^2-11x-6 = (x+1)(x-6)(x^2+x+1)\text{.}\)We should check whether \(x^2+x+1\) is reducible or not. If we try to find its roots with the quadratic equation, we get \(\dfrac{-1\pm\sqrt{-3}}{2}\text{,}\) which are not real numbers. So, we’re at the end of our factoring.
-
Without seeing any obvious patterns, we start hunting for roots. Since we have all integer coefficients, if there are any integer roots, they will divide our constant term, \(-15\text{.}\) So, our candidates for roots are \(\pm1,\,\pm3,\,\pm5,\) and \(\pm15\text{.}\) Write \(f(x) = 2x^4 + 12x^3 - x^2 - 52x + 15\text{.}\)\begin{align*} f(-1)&\neq 0& \color{red}{f(-3)}&\color{red}{=0}& \color{red}{f(-5)}&\color{red}{=0}& f(-5)&\neq 0\\ f(1)&\neq 0& f(3)&\neq 0& f(5)&\neq 0& f(15)&\neq 0& \end{align*}Since \(x=-3\) and \(x=-5\) are roots of our polynomial, it has factors \((x+3)\) and \((x+5)\text{.}\) Note \((x+3)(x+5) = x^2+8x+15\text{.}\) We use long division to move forward.So, \(2x^4+12x^3-x^2-52x+15= (x+3)(x+5)(2x^2-4x+1)\text{.}\)We should check whether \(2x^2-4x+1\) is reducible or not. There’s not an obvious way to factor it, but we can use the quadratic equation. This gives us roots \(\dfrac{4\pm\sqrt{16-8}}{4}=1\pm\dfrac{\sqrt2}{2}\text{.}\) So, we have two more linear factors.Specifically:\begin{align*} 2x^4+12x^3-x^2-52x+15&= (x+3)(x+5)\left(x-\Big(1+\frac{\sqrt2}{2}\Big)\right) \\ &\hskip1.6in\left(x-\Big(1-\frac{\sqrt2}{2}\Big)\right) \end{align*}
1.10.4.6.
Solution.
1.10.4.7. (✳).
Solution.
1.10.4.8. (✳).
Solution.
1.10.4.9. (✳).
Solution.
1.10.4.10. (✳).
Solution.
1.10.4.11. (✳).
Solution.
- The degree of the numerator \(x-13\) is one, which is strictly smaller than the dergee of the denominator \(x^2-x-6\text{,}\) which is two. So we don’t need long division to pull out a polynomial.
- Next we factor the denominator.\begin{gather*} x^2-x-6 = (x-3)(x+2) \end{gather*}
- Next we find the partial fraction decomposition of the integrand. It is of the form\begin{gather*} \frac{x-13}{(x-3)(x+2)} =\frac{A}{x-3} + \frac{B}{x+2} \end{gather*}To find \(A\) and \(B\text{,}\) using the sneaky method, we cross multiply by the denominator.\begin{gather*} x-13 = A(x+2) + B(x-3) \end{gather*}Now we can find \(A\) by evaluating at \(x=3\)\begin{gather*} 3-13 = A(3+2) + B(3-3) \implies \color{red}{ A=-2} \end{gather*}and find \(B\) by evaluating at \(x=-2\text{.}\)\begin{gather*} -2-13 = A(-2+2) + B(-2-3) \implies \color{red}{ B=3} \end{gather*}(Hmmm. \(A\) and \(B\) are nice round numbers. Sure looks like a rigged exam or homework question.) Our partial fraction decomposition is\begin{gather*} \frac{x-13}{(x-3)(x+2)} =\frac{-2}{x-3} + \frac{3}{x+2} \end{gather*}As a check, we recombine the right hand side and make sure that it matches the left hand side.\begin{gather*} \frac{-2}{x-3} + \frac{3}{x+2} =\frac{-2(x+2)+3(x-3)}{(x-3)(x+2)} =\frac{x-13}{(x-3)(x+2)} \end{gather*}
- Finally, we evaluate the integral.\begin{align*} \int \frac{x-13}{x^2-x-6}\dee{x} &=\int\bigg(\frac{-2}{x-3} + \frac{3}{x+2}\bigg)\dee{x}\\ &=-2\log|x-3|+3\log|x+2|+C \end{align*}
1.10.4.12. (✳).
Solution.
- The degree of the numerator \(5x+1\) is one, which is strictly smaller than the dergee of the denominator \(x^2+5x+6\text{,}\) which is two. So we do not long divide to pull out a polynomial.
- Next we factor the denominator.\begin{gather*} x^2+5x+6 = (x+2)(x+3) \end{gather*}
- Next we find the partial fraction decomposition of the integrand. It is of the form\begin{gather*} \frac{5x+1}{(x+2)(x+3)} =\frac{A}{x+2} + \frac{B}{x+3} \end{gather*}To find \(A\) and \(B\text{,}\) using the sneaky method, we cross multiply by the denominator.\begin{gather*} 5x+1= A(x+3) + B(x+2) \end{gather*}Now we can find \(A\) by evaluating at \(x=-2\)\begin{gather*} -10+1 = A(-2+3) + B(-2+2) \implies\color{red}{ A=-9} \end{gather*}and find \(B\) by evaluating at \(x=-3\text{.}\)\begin{gather*} -15+1 = A(-3+3) + B(-3+2) \implies\color{red}{ B=14} \end{gather*}So our partial fraction decomposition is\begin{gather*} \frac{5x+1}{(x+2)(x+3)} =\frac{-9}{x+2} + \frac{14}{x+3} \end{gather*}As a check, we recombine the right hand side and make sure that it matches the left hand side\begin{gather*} \frac{-9}{x+2} + \frac{14}{x+3} =\frac{-9(x+3)+14(x+2)}{(x+2)(x+3)} =\frac{5x+1}{(x+2)(x+3)} \end{gather*}
- Finally, we evaluate the integral\begin{align*} \int \frac{5x+1}{x^2+5x+6}\dee{x} &=\int\bigg(\frac{-9}{x+2} + \frac{14}{x+3}\bigg)\dee{x}\\ &=-9\log|x+2|+14\log|x+3|+C \end{align*}
1.10.4.13.
Solution.
-
Since the degree of the numerator is the same as the degree of the denominator, we need to pull out a polynomial.That is,\begin{align*} \int \frac{5x^2-3x-1}{x^2-1}\dee{x} &= \int\left(5+ \frac{-3x+4}{x^2-1}\right)\dee{x} \\ &=5x+ \int \frac{-3x+4}{x^2-1}\dee{x} \end{align*}
- Again, there’s no obvious substitution for the new integrand, so we want to use partial fraction. The denominator factors as \((x-1)(x+1)\text{,}\) so our decomposition has this form:\begin{align*} \frac{-3x+4}{x^2-1}&=\frac{-3x+4}{(x-1)(x+1)} = \frac{A}{x-1}+\frac{B}{x+1} \\ &= \frac{(A+B)x+(A-B)}{(x-1)(x+1)} \end{align*}So, (1) \(A+B=-3\) and (2) \(A-B=4\text{.}\)
- We solve (2) for \(A\) in terms of \(B\text{,}\) namely \(A=4+B\text{.}\) Plugging this into (1), we see \((4+B)+B=-3\text{.}\) So, \(\textcolor{red}{B=-\frac{7}{2}}\text{,}\) and \(\textcolor{red}{A=\frac{1}{2}}\text{.}\)
- Now we can write our integral in a friendlier form and evaluate.\begin{align*} \int \frac{5x^2-3x-1}{x^2-1}\dee{x} =&=5x+ \int \frac{-3x+4}{x^2-1}\dee{x} \\ &=5x+\int \frac{1/2}{x-1} - \frac{7/2}{x+1} \dee{x}\\ &=5x+\frac{1}{2}\log|x-1| - \frac{7}{2}\log|x+1|+C \end{align*}
1.10.4.14.
Solution.
- The degree of the numerator is the same as the degree of the denominator. Since it’s not smaller, we need to re-write our integrand. We could do this using long division, but this case is simple enough to do more informally.\begin{align*} \frac{4x^4+14x^2+2}{4x^4+x^2}&=\frac{4x^4+x^2+13x^2+2}{4x^4+x^2}\\ &=\frac{4x^4+x^2}{4x^4+x^2}+\frac{13x^2+2}{4x^4+x^2}\\ &=1+\frac{13x^2+2}{4x^4+x^2} \end{align*}
- The denominator factors as \(x^2(4x^2+1)\text{.}\)
- We want to find the partial fraction decomposition of the fractional part of our simplified integrand.\begin{align*} \frac{13x^2+2}{4x^4+x^2}&= \frac{13x^2+2}{x^2(4x^2+1)} = \frac{A}{x}+\frac{B}{x^2}+\frac{Cx+D}{4x^2+1}\\ \end{align*}
Multiply through by the original denominator.
\begin{align*} 13x^2+2&=Ax(4x^2+1)+B(4x^2+1)+(Cx+D)x^2 \tag{1}\\ \end{align*}Setting \(x=0\) gives us:
\begin{align*} \textcolor{red}{2}&\color{red}{=B}\\ \end{align*}We use \(B=2\) to simplify Equation (1).
\begin{align*} 13x^2+2&=Ax(4x^2+1)+\textcolor{red}{2}(4x^2+1)+(Cx+D)x^2\\ 5x^2&=Ax(4x^2+1)+(Cx+D)x^2\\ 5x&=A(4x^2+1)+(Cx+D)x\tag{2}\\ \end{align*}Again, let \(x=0\text{.}\)
\begin{align*} \color{red}{0}&\color{red}{=A}\\ \end{align*}Using \(A=0\text{,}\) simplify Equation (2).
\begin{align*} 5x&=(Cx+D)x\\ 5&=Cx+D\\ \color{red}{C}& \textcolor{red}{=0},\qquad\color{red}{D=5} \end{align*} - Now we can write our integral in pieces.\begin{align*} \int \frac{4x^4+14x^2+2}{4x^4+x^2} \dee{x}&= \int \left(1+\frac{13x^2+2}{4x^4+x^2}\right) \dee{x}\\ &=\int\left(1+ \frac{\textcolor{red}2}{x^2}+\frac{\textcolor{red}5}{4x^2+1}\right) \dee{x}\\ &=x -\frac{2}{x} + \int \frac{5}{(2x)^2+1} \dee{x}\\ \end{align*}
Substitute \(u=2x\text{,}\) \(\dee{u}=2\dee{x}\text{.}\)
\begin{align*} & =x -\frac{2}{x} + \int \frac{5/2}{u^2+1} \dee{u}\\ &=x-\frac{2}{x}+\frac{5}{2}\arctan u +C\\ &=x-\frac{2}{x}+\frac{5}{2}\arctan (2x) +C \end{align*}
1.10.4.15.
Solution.
- Since the numerator has strictly smaller degree than the denominator, we don’t need to start off with a long division.
- We do, however, need to factor the denominator. We can immediately pull out \(x^2\text{;}\) the remaining part is \(x^2-2x+1 = (x-1)^2\text{.}\)
- Now we can perform our partial fraction decomposition.\begin{align*} \frac{x^2+2x-1}{x^4-2x^3+x^2}&= \frac{x^2+2x-1}{x^2(x-1)^2} = \frac{A}{x}+\frac{B}{x^2}+\frac{C}{x-1}+\frac{D}{(x-1)^2}\\ \end{align*}That is, \(\textcolor{red}{A=C=0}\text{.}\)
Multiply both sides by the original denominator.
\begin{align*} x^2+2x-1&=Ax(x\!-\!1)^2+B(x\!-\!1)^2+Cx^2(x\!-\!1)+Dx^2\tag{1}\\ \end{align*}To be sneaky, we set \(x=0\text{,}\) and find:
\begin{align*} \color{red}{-1}&\color{red}{=B}\\ \end{align*}We also set \(x=1\text{,}\) and find:
\begin{align*} \color{red}{2}&\color{red}{=D}\\ \end{align*}We use \(B\) and \(D\) to simplify Equation (1).
\begin{align*} x^2+2x-1&=Ax(x-1)^2\textcolor{red}{-1}(x-1)^2+Cx^2(x-1)+\textcolor{red}{2}x^2\\ 0&=Ax(x-1)^2+Cx^2(x-1)\\ &=x(x-1)[(A+C)x-A]\\ \mbox{So,}\qquad 0&=(A+C)x-A \end{align*} - Now we can evaluate our integral.\begin{align*} \int \frac{x^2+2x-1}{x^4-2x^3+x^2} \dee{x}&=\int \left(\frac{-1}{x^2}+\frac{2}{(x-1)^2}\right) \dee{x}\\ &=\frac{1}{x}-\frac{2}{x-1}+C \end{align*}
1.10.4.16.
Solution.
- The degree of the numerator is less than the degree of the denominator.
- We need to factor the denominator. The first two terms have the same ratio as the last two terms.\begin{align*} 2x^3-x^2-8x+4&=x^2(2x-1)-4(2x-1)\\ &= (x^2-4)(2x-1)\\ &=(x-2)(x+2)(2x-1) \end{align*}
- Now we find our partial fraction decomposition.\begin{align*} \frac{ 3x^2-4x-10}{2x^3-x^2-8x+4}&=\frac{ 3x^2-4x-10}{(x-2)(x+2)(2x-1)}\\ &=\frac{A}{x-2}+\frac{B}{x+2}+\frac{C}{2x-1}\\ \end{align*}
Multiply both sides by the original denominator.
\begin{align*} 3x^2-4x-10&=A(x+2)(2x-1)+B(x-2)(2x-1)\\ &\hskip2in+C(x-2)(x+2)\\ \end{align*}Distinct linear factors is the best possible scenario for the sneaky method. First, let’s set \(x=2\text{.}\)
\begin{align*} 3(4)-4(2)-10&=A(4)(3)+B(0)+C(0)\\ \color{red}{A}&\color{red}{=-\frac{1}{2}}\\ \end{align*}Now, let \(x=-2\text{.}\)
\begin{align*} 3(4)-4(-2)-10&=A(0)+B(-4)(-5)+C(0)\\ \color{red}{B}&\color{red}{=\frac{1}{2}}\\ \end{align*}Finally, let \(x=\frac{1}{2}\text{.}\)
\begin{align*} \frac{3}{4}-2-10&=A(0)+B(0)+C\left(-\frac{3}{2}\right)\left(\frac{5}{2}\right)\\ \color{red}{C}&\color{red}{=3} \end{align*} - Now we can evaluate our integral in its new form.\begin{align*} &\int \frac{ 3x^2-4x-10}{2x^3-x^2-8x+4} \dee{x}=\int \left(\frac{-1/2}{x-2}+\frac{1/2}{x+2}+\frac{3}{2x-1}\right) \dee{x}\\ &\hskip0.5in=-\frac{1}{2}\log|x-2| + \frac{1}{2}\log|x+2| + \frac{3}{2}\log|2x-1|+C\\ &\hskip0.5in=\frac{1}{2}\log\left| \frac{x+2}{x-2} \right| + \frac{3}{2}\log|2x-1|+C \end{align*}
1.10.4.17.
Solution.
- The numerator has smaller degree than the denominator.
-
We need to factor the denominator. In the absence of any clues, we look for an integer root. The constant term is 5, so the possible integer roots are \(\pm 1\) and \(\pm 5\text{.}\) Name \(f(x) = 2x^3+11x^2+6x+5\text{.}\)\begin{equation*} f(-1)\neq0 \qquad {\color{red}{f(-5)=0}} \qquad f(1)\neq 0 \qquad f(5)\neq 0 \end{equation*}So, \((x+5)\) is a factor of the denominator.
-
We use long division to pull out the factor of \((x+5)\text{.}\)That is, our denominator is \((x+5)(2x^2+x+1)\text{.}\)
- The quadratic function \(2x^2+x+1\) is irreducible: we can see this by using the quadratic equation, and finding no real roots. So, we are ready to find our partial fraction decomposition.\begin{align*} \frac{10x^2+24x+8}{2x^3+11x^2+6x+5}&=\frac{10x^2+24x+8}{(x+5)(2x^2+x+1)}\\ &=\frac{A}{x+5}+\frac{Bx+C}{2x^2+x+1}\\ \end{align*}So, \(\textcolor{red}{B=4}\) and \(\textcolor{red}{C=1}\text{.}\)
Multiply through by the original denominator.
\begin{align*} 10x^2+24x+8&=A(2x^2+x+1)+(Bx+C)(x+5)\tag{1}\\ \end{align*}Set \(x=-5\text{.}\)
\begin{align*} 10(25)-24(5)+8&=A(2(25)-5+1) + (B(-5)+C)(0)\\ \color{red}{A}&\color{red}{=3}\\ \end{align*}Using our value of \(A\text{,}\) we simplify Equation (1).
\begin{align*} 10x^2+24x+8&=\textcolor{red}{3}(2x^2+x+1)+(Bx+C)(x+5)\\ 4x^2+21x+5&=(Bx+C)(x+5)\\ \end{align*}We factor the left side. We know \((x+5)\) must be one of its factors.
\begin{align*} (4x+1)(x+5)&=(Bx+C)(x+5)\\ 4x+1&=Bx+C \end{align*} - Now we can write our integral in smaller pieces.\begin{align*} \int_0^1 \frac{10x^2+24x+8}{2x^3+11x^2+6x+5} \dee{x}&= \int_0^1 \left( \frac{3}{x+5} + \frac{4x+1}{2x^2+x+1}\right) \dee{x}\\ \end{align*}
The antiderivative of the left fraction is \(3\log|x+5|\text{.}\) For the right fraction, we use the substitution \(u=2x^2+x+1\text{,}\) \(\dee{u}=(4x+1)\dee{x}\) to antidifferentiate.
\begin{align*} &=\big[3\log|x+5| + \log|2x^2+x+1|\big]_0^1\\ &=3\log 6 + \log 4 - 3\log 5 -\log 1\\ &=\log \left(\frac{4\cdot 6^3}{5^3}\right) \end{align*}
1.10.4.18.
Solution.
Let \(u=\cos x\text{,}\) \(\dee{u}=-\sin x \dee{x}\text{.}\)
\begin{align*} &=\int\frac{-1}{1-u^2} \dee{u} = \int\frac{-1}{(1+u)(1-u)} \dee{u}\\ \end{align*}We see an opportunity for partial fraction.
\begin{align*} \frac{-1}{(1+u)(1-u)}&=\frac{A}{1+u}+\frac{B}{1-u}\\ \end{align*}Multiply both sides by the original denominator.
\begin{align*} -1&=A(1-u)+B(1+u)\\ \end{align*}Let \(u=1\text{.}\)
\begin{align*} -1&=2B \qquad\Rightarrow\color{red}{ B = -\frac{1}{2}}\\ \end{align*}Let \(u=-1\text{.}\)
\begin{align*} -1&=2A \qquad\Rightarrow \color{red}{A = -\frac{1}{2}}\\ \end{align*}We can now re-write our integral.
\begin{align*} \int \csc x \dee{x}&= \int\frac{-1}{(1+u)(1-u)} \dee{u} =\int \left(\frac{-1/2}{1+u} + \frac{-1/2}{1-u}\right) \dee{u}\\ &=-\frac{1}{2}\log|1+u| +\frac{1}{2}\log|1-u|+C\\ &=\frac{1}{2}\log\left| \frac{1-u}{1+u}\right|+C\\ &=\frac{1}{2}\log\left| \frac{1-\cos x}{1+\cos x}\right|+C \end{align*}1.10.4.19.
Solution.
Let \(u=\cos x\text{,}\) \(\dee{u}=-\sin x \dee{x}\text{.}\)
\begin{align*} &=\int\frac{-1}{(1-u^2)^2} \dee{u} \end{align*}1.10.4.20.
Solution.
- First, we check that the numerator has strictly smaller degree than the denominator, so we don’t have to use long division.
- Second, we factor the denominator. We can immediately pull out a factor of \(x^2\text{;}\) then we’re left with the quadratic polynomial \(x^2+5x+10\text{.}\) Using the quadratic equation, we check that this has no real roots, so it is irreducible.
- Once we know the factorization of the denominator, we can set up our decomposition.\begin{align*} \frac{3x^3+15x^2+35x+10}{x^4+5x^3+10x^2}&= \frac{3x^3+15x^2+35x+10}{x^2(x^2+5x+10)}\\ &=\frac{A}{x}+\frac{B}{x^2}+\frac{Cx+D}{x^2+5x+10} \end{align*}We multiply both sides by the original denominator.\begin{align*} &3x^3+15x^2+35x+10\\ &\hskip0.1in=Ax(x^2+5x+10) + B(x^2+5x+10)+(Cx+D)x^2 \tag{1} \end{align*}Following the “Sneaky Method,” we plug in \(x=0\text{.}\)\begin{align*} 0+10&=A(0)+B(10)+(C(0)+D)(0)\\ \color{red}{B}&\color{red}{=1} \end{align*}
- Knowing \(B\) allows us to simplify our Equation (1).\begin{align*} 3x^3+15x^2+35x+10&=Ax(x^2+5x+10) + \textcolor{red}{1}(x^2+5x+10)\\ &\hskip2in+(Cx+D)x^2\\ 3x^3+14x^2+30x&=Ax(x^2+5x+10) +(Cx+D)x^2 \end{align*}We can factor \(x\) out of both sides of the equation.\begin{align*} 3x^2+14x+30&=A(x^2+5x+10) +(Cx+D)x \tag{2} \end{align*}
- Again, we set \(x=0\text{.}\)\begin{align*} 0+30&=A(10) +(C(0+D)(0)\\ \color{red}{A}&\color{red}{=3} \end{align*}
- We simplify Equation (2), using \(A=3\text{.}\)\begin{align*} 3x^2+14x+30&=\textcolor{red}{3}(x^2+5x+10) +(Cx+D)x\\ -x&=Cx^2+Dx\\ \color{red}{C}&\color{red}{=0,\quad D=-1} \end{align*}
- Now that we have our coefficients, we can re-write our integral in a friendlier form.\begin{align*} &\int_1^2 \frac{3x^3+15x^2+35x+10}{x^4+5x+10x^2} \dee{x}\\ &\hskip0.5in=\int_1^2 \left(\frac{3}{x}+\frac{1}{x^2}-\frac{1}{x^2+5x+10} \right) \dee{x}\\ &\hskip0.5in=\left[3\log|x|-\frac{1}{x}\right]_1^2-\int_1^2\frac{1}{x^2+5x+10} \dee{x}\\ &\hskip0.5in=3\log 2+\frac{1}{2}-\int_1^2\frac{1}{x^2+5x+10} \dee{x} \end{align*}The remaining integral is the reciprocal of a quadratic polynomial, much like \(\dfrac{1}{1+x^2}\text{,}\) whose antiderivative is arctangent. We complete the square and use the substitution \(u=\left(\frac{2x+5}{\sqrt{15}}\right)\text{,}\) \(\dee{u}=\frac{2}{\sqrt{15}} \dee{x}\text{.}\)\begin{align*} \int_1^2 \frac{1}{x^2+5x+10} \dee{x}&=\int_1^2 \frac{1}{\left(x+\frac{5}{2}\right)^2+\frac{15}{4}} \dee{x}\\ &=\frac{4}{15}\int_1^2 \frac{1}{\left(\frac{2x+5}{\sqrt{15}}\right)^2+1} \dee{x}\\ &=\frac{2}{\sqrt{15}}\int_{7/\sqrt{15}}^{9/\sqrt{15}} \frac{1}{u^2+1} \dee{u}\\ &=\frac{2}{\sqrt{15}}\Big[\arctan u\Big]_{7/\sqrt{15}}^{9/\sqrt{15}}\\ &=\frac{2}{\sqrt{15}}\left(\!\arctan\left(\frac{9}{\sqrt{15}}\right) \!-\!\arctan\left(\frac{7}{\sqrt{15}}\!\right)\right) \end{align*}So, all together,\begin{align*} &\int_1^2 \frac{3x^3+15x^2+35x+10}{x^4+5x^3+10x^2} \dee{x}\\ &\hskip0.5in=3\log 2 + \frac{1}{2}-\frac{2}{\sqrt{15}}\bigg(\arctan\left(\frac{9}{\sqrt{15}}\right)\\ &\hskip1.5in-\arctan\left(\frac{7}{\sqrt{15}}\right)\bigg) \end{align*}
1.10.4.21.
Solution.
-
First piece: \(\int \frac{3}{x^2+2} \dee{x}\text{.}\) The fraction looks somewhat like the derivative of arctangent, so we can massage it to find an appropriate substitution.\begin{align*} \int \frac{3}{x^2+2} \dee{x}&=\frac{3}{2}\int \frac{1}{\left(\frac{x}{\sqrt2}\right)^2+1} \dee{x}\\ \end{align*}
Use the substitution \(u=\frac{x}{\sqrt2}\text{,}\) \(\dee{u} = \frac{1}{\sqrt{2}} \dee{x}.\)
\begin{align*} &=\frac{3}{\sqrt2}\int\frac{1}{u^2+1} \dee{u}\\ &=\frac{3}{\sqrt2}\arctan u+C\\ &=\frac{3}{\sqrt2}\arctan \left(\frac{x}{\sqrt2}\right)+C \end{align*} -
The next piece is \(\int \frac{x-3}{(x^2+2)^2} \dee{x}\text{.}\) If the numerator were only \(x\) (and no constant), we could use the substitution \(u=x^2+2\text{,}\) \(\dee{u} = 2x \dee{x}\text{.}\) So, to that end, we can break up that fraction into \(\frac{x}{(x^2+2)^2}-\frac{3}{(x^2+2)^2}\text{.}\) For now, we only evaluate the first half.\begin{align*} \int \frac{x}{(x^2+2)^2} \dee{x}&=\frac{1}{2}\int\frac{1}{u^2} \dee{u} = -\frac{1}{2u}+C\\ &=-\frac{1}{2x^2+4}+C \end{align*}
-
That leaves us with the final piece, \(\frac{3}{(x^2+2)^2}\text{,}\) which is the hardest. We saw something similar in Question 1.9.2.20 in Section 1.9: we can use the substitution \(x=\sqrt{2}\tan\theta\text{,}\) \(\dee{x} = \sqrt{2}\sec^2\theta \dee{\theta}\text{.}\)\begin{align*} \int\frac{3}{(x^2+2)^2} \dee{x}&=\int\frac{3}{(2\tan^2\theta+2)^2}\sqrt{2}\sec^2\theta \dee{\theta}\\ &=\int\frac{3}{4\sec^4\theta}\sqrt{2}\sec^2\theta \dee{\theta}\\ &=\frac{3}{2\sqrt{2}}\int\cos^2\theta \dee{\theta}\\ &=\frac{3}{4\sqrt{2}}\int\big(1+\cos(2\theta)\big) \dee{\theta}\\ &=\frac{3}{4\sqrt{2}}\left(\theta+\frac{1}{2}\sin(2\theta)\right)+C\\ &=\frac{3}{4\sqrt{2}}\left(\theta+\sin\theta\cos\theta\right)+C\\ &=\frac{3}{4\sqrt{2}}\left(\arctan\left(\frac{x}{\sqrt2}\right)+\frac{x\sqrt{2}}{x^2+2}\right)+C \end{align*}From our substitution, \(\tan\theta = \frac{x}{\sqrt2}\text{.}\) So, we can draw a right triangle with angle \(\theta\text{,}\) opposite side \(x\text{,}\) and adjacent side \(\sqrt2\text{.}\) Then by the Pythagorean Theorem, the hypotenuse has length \(\sqrt{x^2+2}\text{,}\) and this gives us \(\sin\theta\) and \(\cos\theta\text{.}\)
1.10.4.22.
Solution.
1.10.4.23.
Solution.
For the first integral, use the substitution \(w=\frac{x}{\sqrt5}\text{,}\) \(\dee{w} = \frac{1}{\sqrt5} \dee{x}\text{.}\)
\begin{align*} &= \frac{3}{2}x^2+\frac{1}{\sqrt{5}}\int\frac{1}{w^2+1}\dee{w}+ \frac{3}{2}\log|u|-\frac{3}{2u}\\ &= \frac{3}{2}x^2+\frac{1}{\sqrt{5}}\arctan w+ \frac{3}{2}\log|x^2+5|-\frac{3}{2x^2+10}+C\\ &= \frac{3}{2}x^2+\frac{1}{\sqrt{5}}\arctan \left(\frac{x}{\sqrt5}\right)+ \frac{3}{2}\log|x^2+5|-\frac{3}{2x^2+10}+C \end{align*}1.10.4.24.
Solution.
We use the substitution \(x=\sin \theta\text{,}\) \(\dee{x}=\cos\theta \dee{\theta}\text{.}\)
\begin{align*} &=\int \frac{1}{3x-x^2-2} \dee{x}=\int \frac{-1}{x^2-3x+2} \dee{x} =\int \frac{-1}{(x-1)(x-2)} \dee{x}\\ \end{align*}Now we can find a partial fraction decomposition.
\begin{align*} \frac{-1}{(x-1)(x-2)}&=\frac{A}{x-1}+\frac{B}{x-2}\\ -1&=A(x-2)+B(x-1)\\ \end{align*}Setting \(x=1\) and \(x=2\text{,}\) we see
\begin{align*} \color{red}{A}&\color{red}{=1},\quad \color{red}{B=-1} \end{align*}1.10.4.25.
Solution.
The factor \(x^2+x+1\) is an irreducible quadratic, so the denominator is completely factored. Now we can use partial fraction decomposition.
\begin{align*} \frac{1}{x\left(x^2+x+1\right)}&=\frac{A}{x}+\frac{Bx+C}{x^2+x+1}\\ 1&=A(x^2+x+1)+(Bx+C)x\\ 1&=(A+B)x^2+(A+C)x+A \end{align*}In step (\(*\)), we set ourselves up so that we could evaluate the second integral with the substitution \(u=x^2+x+1\text{.}\) For the remaining integral, we complete the square, so that the integrand looks something like the derivative of arctangent.
\begin{align*} &=\log|x| - \frac{1}{2}\log|x^2+x+1|-\int \frac{1/2}{\left(x+\frac{1}{2}\right)^2+\frac{3}{4}} \dee{x}\\ &=\log|x| - \frac{1}{2}\log|x^2+x+1|-\frac{2}{3}\int \frac{1}{\left(\frac{2x+1}{\sqrt3}\right)^2+1} \dee{x}\\ \end{align*}We use the substitution \(u=\frac{2x+1}{\sqrt{3}}\text{,}\) \(\dee{u} = \frac{2}{\sqrt3}\text{.}\)
\begin{align*} &=\log|x| - \frac{1}{2}\log|x^2+x+1|-\frac{1}{\sqrt3}\int \frac{1}{u^2+1} \dee{u}\\ &=\log|x| - \frac{1}{2}\log|x^2+x+1|-\frac{1}{\sqrt3}\arctan u+C\\ &=\log|x| - \frac{1}{2}\log|x^2+x+1|-\frac{1}{\sqrt3}\arctan \left(\frac{2x+1}{\sqrt3}\right)+C\\ &=\log|e^t| - \frac{1}{2}\log|e^{2t}+e^t+1|-\frac{1}{\sqrt3}\arctan \left(\frac{2e^t+1}{\sqrt3}\right)+C\\ &=t - \frac{1}{2}\log|e^{2t}+e^t+1|-\frac{1}{\sqrt3}\arctan \left(\frac{2e^t+1}{\sqrt3}\right)+C \end{align*}1.10.4.26.
Solution.
-
Solution 1: We use the substitution \(u=\sqrt{1+e^x}\text{.}\)Then \(\dee{u} = \dfrac{e^x}{2\sqrt{1+e^x}}\dee{x}\text{,}\) so \(\dee{x}=\dfrac{2u}{u^2-1}\dee{u}\text{.}\)\begin{align*} \int\sqrt{1+e^x} \dee{x}&=\int u\cdot\dfrac{2u}{u^2-1}\dee{u} = \int \dfrac{2u^2}{u^2-1}\dee{u}\\ &= \int \dfrac{2(u^2-1)+2}{u^2-1}\dee{u} = \int\left(2+ \dfrac{2}{u^2-1}\right)\dee{u} \end{align*}We use a partial fraction decomposition on the fractional part of the integrand.\begin{align*} \frac{2}{u^2-1} &= \frac{2}{(u-1)(u+1)}=\frac{A}{u-1}+\frac{B}{u+1}\\ &=\frac{(A+B)u+(A-B)}{(u-1)(u+1)} \end{align*}For the right hand side to match the left hand side, we need\begin{gather*} A+B=0,\quad A-B=2 \implies \color{red}{A=1,\quad B=-1} \end{gather*}So the integral\begin{align*} &\int\sqrt{1+e^x} \dee{x}=\int\left(2+ \dfrac{2}{u^2-1}\right)\dee{u} \\ &\hskip0.25in=\int\left( 2 + \frac{1}{u-1} - \frac{1}{u+1}\right) \dee{u}\\ &\hskip0.25in=2u+\log|u-1|-\log|u+1|+C=2u+\log\left| \frac{u-1}{u+1}\right|+C\\ &\hskip0.25in=2\sqrt{1+e^x}+\log\left| \frac{\sqrt{1+e^x}-1}{\sqrt{1+e^x}+1}\right|+C \end{align*}
- Solution 2: It might not occur to us right away to use the fruitful substitution in Solution 1. More realistically, we might start with the “inside function,” \(u=1+e^x\text{.}\) Then \(\dee{u}=e^x\,\dee{x}\text{,}\) so \(\dee{x} = \frac{1}{u-1}\dee{u}\text{.}\)\begin{align*} \int\sqrt{1+e^x} \dee{x}&=\int\frac{ \sqrt{u}}{u-1}\dee{u}\\ \end{align*}Now we can use partial fraction decomposition.
This isn’t quite a rational function, because we have a square root on top. If we could turn it into a rational function, we could use partial fraction. To that end, let \(w=\sqrt{u}\text{,}\) \(\dee{w} = \frac{1}{2\sqrt{u}}\dee{u}\text{,}\) so \(\dee{u}=2w\dee{w}\text{.}\)
\begin{align*} &=\int \frac{w}{w^2-1}2w\dee{w}=\int \frac{2w^2}{w^2-1}\dee{w}\\ &=\int \frac{2(w^2-1)+2}{w^2-1}\dee{w} = \int 2 + \frac{2}{w^2-1}\dee{w} \end{align*}\begin{align*} \frac{2}{w^2-1} &= \frac{2}{(w-1)(w+1)} = \frac{A}{w-1}+\frac{B}{w+1} \\ &= \frac{(A+B)w+(A-B)}{(w-1)(w+1)} \end{align*}For the left and right hand sides to match, we need\begin{gather*} A+B=0,\quad A-B=2 \implies \color{red}{ A=1,\quad B=-1} \end{gather*}This allows us to antidifferentiate.\begin{align*} \int\sqrt{1+e^x} \dee{x}&= \int \left(2 + \frac{2}{w^2-1}\right)\dee{w}\\ &=\int \left(2 + \frac{1}{w-1}-\frac{1}{w+1}\right) \dee{w}\\ &= 2w + \log|w-1| - \log|w+1|+C\\ &=2w+\log\left| \frac{w-1}{w+1}\right|+C\\ &=2\sqrt{u}+\log\left| \frac{\sqrt{u}-1}{\sqrt{u}+1}\right|+C\\ &=2\sqrt{1+e^x}+\log\left| \frac{\sqrt{1+e^x}-1}{\sqrt{1+e^x}+1}\right|+C \end{align*}
1.10.4.27. (✳).
Solution.
- We cut \(\cR\) into thin horizontal strips of width \(\dee{y}\) as in the figure on the right above.
- When we rotate \(\cR\) about the \(y\)-axis, each strip sweeps out a thin washer
- whose outer radius is \(r_{out}=4\text{,}\) and
- whose inner radius is \(r_{in}= \sqrt{25-\frac{100}{y^2}}\) when \(y\ge \frac{10}{\sqrt{25-3^2}} = \frac{10}{4} =\frac{5}{2}\) (see the red strip in the figure on the right above), and whose inner radius is \(r_{in}= 3\) when \(y\le \frac{5}{2}\) (see the blue strip in the figure on the right above) and
- whose thickness is \(\dee{y}\) and hence
- whose volume is \(\pi(r_{out}^2 - r_{in}^2)\dee{y} = \pi\big(\frac{100}{y^2}-9\big)\dee{y}\) when \(y\ge \frac{5}{2}\) and whose volume is \(\pi(r_{out}^2 - r_{in}^2)\dee{y} = 7 \pi\,\dee{y}\) when \(y\le \frac{5}{2}\) and
- As our bottommost strip is at \(y=0\) and our topmost strip is at \(y=\frac{10}{3}\) (since at the top \(x=4\) and \(y= \frac{10}{\sqrt{25-x^2}} =\frac{10}{\sqrt{25-4^2}} =\frac{10}{3}\)), the volume is\begin{align*} &\int _{5/2}^{10/3} \pi\Big(\frac{100}{y^2}-9\Big)\dee{y} +\int _ 0^{5/2}7 \pi\,\dee{y}\\ &=\pi{\Big[-\frac{100}{y}-9y\Big]}_{5/2}^{10/3} +\frac{35}{2}\pi\\ &=\pi \Big[-30+40-30+\frac{45}{2}\Big] +\frac{35}{2}\pi\\ &=20\pi \end{align*}
1.10.4.28.
Solution.
\(x\) | \(\frac{4}{3+x^2}\) | \(\frac{2}{x(x+1)}\) | Top: |
\(1/2\) | \(16/13\) | \(8/3\) | \(\frac{2}{x(x+1)}\) |
\(2\) | \(4/7\) | \(1/3\) | \(\frac{4}{3+x^2}\) |
1.10.4.29.
Solution.
1.11 Numerical Integration
1.11.6 Exercises
1.11.6.1.
Solution.
1.11.6.2.
Solution.
1.11.6.3.
Solution.
-
Differentiating, we find \(f''(x) = -x^2+7x-6\text{.}\) Since \(f''(x)\) is quadratic, we have a pretty good idea of what it looks like.
- It factors as \(f(x) = -(x-6)(x-1)\text{,}\) so its two roots are at \(x=6\) and \(x=1\text{.}\)
- The “flat part” of the parabola is at \(x=3.5\) (since this is exactly half way between \(x=1\) and \(x=6\text{;}\) alternately, we can check that \(f'''(3.5)=0\)).
- Since the coefficient of \(x^2\) is negative, \(f(x)\) is increasing from \(-\infty\) to \(3.5\text{,}\) then decreasing from \(3.5\) to \(\infty\text{.}\)
Therefore, over the interval \([1,6]\text{,}\) the largest positive value of \(f''(x)\) occurs when \(x=3.5\text{,}\) and this is \(f''(3.5) = -(3.5-6)(3.5-1)=6.25\text{.}\)So, we take \(M=6.25\text{.}\) - We differentiate further to find \(f^{(4)}(x)=-2\text{.}\) This is constant everywhere, so we take \(L=|-2|=2\text{.}\)
1.11.6.4.
Solution.
1.11.6.5.
Solution.
-
Let \(f(x) = \cos x\text{.}\) Then \(f^{(4)}(x)=\cos x\text{,}\) so \(|f^{(4)}(x)| \leq 1\) when \(-\pi \leq x \leq \pi\text{.}\) So, using \(L=1\text{,}\) we find the upper bound of the error using Simpson’s rule with \(n=4\) is:\begin{equation*} \frac{L(b-a)^5}{180 n^4} = \frac{(2\pi)^5}{180\cdot 4^4} = \frac{\pi^5}{180\cdot8}\approx 0.2 \end{equation*}The error bound comes from Theorem 1.11.13 in the text. We used a calculator to find the approximate decimal value.
- We use the general form of Simpson’s rule (Equation 1.11.9 in the text) with \(\Delta x = \frac{b-a}{n}=\frac{2\pi}{4} = \frac{\pi}{2}\text{.}\)\begin{align*} A& \approx \frac{\Delta x}{3}\left(f(x_0) + 4f(x_1)+2f(x_2)+4f(x_3)+f(x_4)\right)\\ &=\frac{\pi/2}{3}\left(f(-\pi) + 4f(\tfrac{-\pi}{2})+2f(0)+4f(\tfrac{\pi}{2})+f(\pi)\right)\\ &= \frac{\pi}{6}\left(-1 + 4(0)+2(1)+4(0)-1\right)=0 \end{align*}
- To find the actual error in our approximation, we compare the approximation from (b) to the exact value of \(A\text{.}\) In fact, \(A=0\text{:}\) this is a fact you’ve probably seen before by considering the symmetry of cosine, but it’s easy enough to calculate:\begin{equation*} A = \int_{-\pi}^{\pi} \cos x \dee{x}= \sin \pi - \sin (-\pi) = 0 \end{equation*}So, our approximation was exactly the same as our exact value. The absolute error is 0.
1.11.6.6.
Solution.
for some constants \(C\) and \(D\text{.}\) Now we can find the exact and approximate values of \(\displaystyle\int_0^1 f(x) \dee{x}\text{.}\)
\begin{align*} &\mbox{Exact:}&\int_0^1 f(x) \dee{x}&=\int_0^1 \left(\frac{3}{2}x^2+Cx+D\right) \dee{x}\\ &&&=\left[\frac{1}{2}x^3+\frac{C}{2}x^2+Dx\right]_0^1\\ &&&=\textcolor{blue}{\frac{1}{2}+\frac{C}{2}+D}\\ &\mbox{Approximate:}&\int_0^1 f(x) \dee{x}&\approx \Delta x \left[\frac{1}{2}f(0)+f(\tfrac{1}{2})+\frac{1}{2}f(1)\right]\\ &&&=\frac{1}{2}\bigg[\frac{1}{2}(D)+\left(\frac{3}{8}+\frac{C}{2}+D\right)\\ &&&\hskip1.0in+\frac{1}{2}\left(\frac{3}{2}+C+D\right)\bigg]\\ &&&=\frac{1}{2}\left[\frac{9}{8}+C+2D\right]\\ &&&=\textcolor{red}{\frac{9}{16}+\frac{C}{2}+D} \end{align*}1.11.6.7.
Solution.
16
1.11.6.8.
Solution.
1.11.6.9. (✳).
Solution.
1.11.6.10.
Solution.
1.11.6.11.
Solution.
- For all three approximations, \(\Delta x = \dfrac{b-a}{n}=\dfrac{30-0}{6}=5\text{.}\)
-
For the trapezoidal rule and Simpson’s rule, the \(x\)-values where we evaluate \(\dfrac{1}{x^3+1}\) start at \(x=a=0\) and move up by \(\Delta x = 5\text{:}\) \(x_0=0\text{,}\) \(x_1=5\text{,}\) \(x_2=10\text{,}\) \(x_3=15\text{,}\) \(x_4=20\text{,}\) \(x_5=25\text{,}\) and \(x_6=30\text{.}\)
-
For the midpoint rule, the \(x\)-values where we evaluate \(\dfrac{1}{x^3+1}\) start at \(x=2.5 = \frac{x_0+x_1}{2}\) and move up by \(\Delta x = 5\text{:}\) \(\bar x_1=2.5\text{,}\) \(\bar x_2=7.5\text{,}\) \(\bar x_3=12.5\text{,}\) \(\bar x_4=17.5\text{,}\) \(\bar x_5=22.5\text{,}\) and \(\bar x_6=27.5\text{.}\)
- Following Equation 1.11.2, the midpoint rule approximation is:\begin{align*} &\int_0^{30} \frac{1}{x^3+1}\,\dee{x}\approx\Big[f(\bar x_1)+f(\bar x_2)+\cdots +f(\bar x_n)\Big]\De x\\ &=\bigg[\tfrac{1}{\left(2.5\right)^3+1} +\frac{1}{\left(7.5\right)^3+1} +\frac{1}{\left(12.5\right)^3+1} +\frac{1}{\left(17.5\right)^3+1}\\ &\hskip2.25in +\frac{1}{\left(22.5\right)^3+1} +\frac{1}{\left(27.5\right)^3+1} \bigg]5 \end{align*}
- Following Equation 1.11.6, the trapezoidal rule approximation is:\begin{align*} &\int_0^{30} \frac{1}{x^3+1}\,\dee{x}\\ &\approx\Big[\half f(x_0)+f(x_1)+f(x_2)+\cdots+ f(x_{n-1})+\half f(x_n)\Big]\De x\\ &=\bigg[ \frac{1/2}{0^3+1}+ \frac{1}{5^3+1}+ \frac{1}{10^3+1}+ \frac{1}{15^3+1}+ \frac{1}{20^3+1}\\ &\hskip2.5in +\frac{1}{25^3+1}+ \frac{1/2}{30^3+1} \bigg]5 \end{align*}
- Following Equation 1.11.9, the Simpson’s rule approximation is:\begin{align*} &\int_0^{30} \frac{1}{x^3+1}\,\dee{x}\\ &\approx\Big[f(x_0)\!+\!4f(x_1)\!+\!2f(x_2)\!+\!4f(x_3)\!+\!2f(x_4)\!+\!4f(x_{5})\!+\!f(x_6)\Big]\frac{\De x}{3}\\ &=\bigg[\frac{1}{{0}^3+1}\!+\frac{4}{{5}^3+1}\!+\frac{2}{{10}^3+1}\!+\frac{4}{{15}^3+1}\!+\frac{2}{{20}^3+1}\\ &\hskip2.5in+\frac{4}{{25}^3+1}\!+ \frac{1}{{30}^3+1}\bigg]\frac{5}{3} \end{align*}
1.11.6.12. (✳).
Solution.
1.11.6.13. (✳).
Solution.
- We are told that the pancake at height \(x\) is a circular disk of diameter \(f(x)\) and so
- has cross-sectional area \(\pi\big(\frac{f(x)}{2}\big)^2\) and thickness \(\dee{x}\) and hence
- has volume \(\pi\big(\frac{f(x)}{2}\big)^2\dee{x}\text{.}\)
1.11.6.14. (✳).
Solution.
- radius \(\frac{f(x)}{2}\text{,}\)
- width \(\dee{x}\text{,}\) and so
- volume \(\pi\left(\frac{f(x)}{2}\right)^2 \dee{x}=\frac{\pi}{4}f(x)^2 \dee{x}\text{.}\)
1.11.6.15. (✳).
Solution.
1.11.6.16. (✳).
Solution.
1.11.6.17. (✳).
Solution.
1.11.6.18. (✳).
Solution.
1.11.6.19. (✳).
Solution.
17
1.11.6.20. (✳).
Solution.
1.11.6.21. (✳).
Solution.
1.11.6.22. (✳).
Solution.
1.11.6.23. (✳).
Solution.
1.11.6.24. (✳).
Solution.
- \begin{align*} T_4&={\De x}\left[\frac{1}{2}s(0)+s(2)+s(4)+s(6)+\frac{1}{2}s(8)\right]\\ &=2\left[\frac{1.00664}{2}+1.00543+ 1.00435+1.00331+ \frac{1.00233}{2}\right]\\ &=8.03515\\ S_4&=\frac{\De x}{3}\big[s(0)+4s(2)+2s(4)+4s(6)+s(8)\big]\\ &=\frac{2}{3}\big[1.00664+4\times 1.00543+2\times 1.00435+4\times1.00331\\ &\hskip3in + 1.00233\big]\\ &\approx 8.03509 \end{align*}
-
The information \(\big|s^{(k)}(x)\big|\le \dfrac{k}{1000}\text{,}\) with \(k=2\text{,}\) tells us \(|s''(x)|\leq \frac{2}{1000}\) for all \(x\) in the interval \([0,8]\text{.}\) So, we take \(K_2\) (also called \(M\) in your text) to be \(\frac{2}{1000}\text{.}\)Then the absolute error associated with our trapezoid rule approximation is at most\begin{align*} \bigg|\int_a^b f(x)\ \dee{x} -T_n\bigg|&\le \frac{K_2(b-a)^3}{12n^2} \le \frac{2}{1000}\cdot\frac{8^3}{12(4)^2} \le 0.00533 \end{align*}For \(k=4\text{,}\) we see \(|s^{(4)}(x)|\leq \frac{4}{1000}\) for all \(x\) in the interval \([0,8]\text{.}\) So, we take \(K_4\) (also called \(L\) in your text) to be \(\frac{4}{1000}\text{.}\)Then the absolute error associated with our Simpson’s rule approximation is at most\begin{align*} \bigg|\int_a^b f(x)\ \dee{x} -S_n\bigg|&\le \frac{K_4(b-a)^5}{180n^4} \le \frac{4}{1000}\cdot\frac{8^5}{180(4)^4} \le0.00284 \end{align*}
1.11.6.25. (✳).
Solution.
1.11.6.26. (✳).
Solution.
1.11.6.27. (✳).
Solution.
1.11.6.28. (✳).
Solution.
So, we want an even number \(n\) such that
\begin{align*} \frac{4}{30n^4} &\leq 0.00001 = \frac{1}{10^5}\\ n^4 & \geq \frac{40000}{3}\\ n&\geq \sqrt[4]{\frac{40000}{3}}\approx 10.7 \end{align*}1.11.6.29. (✳).
Solution.
1.11.6.30. (✳).
Solution.
We want an integer \(n\) such that
\begin{align*} \frac{1}{3n^2}&\le 0.000005\\ n^2&\ge \frac{4}{12\times 0.000005}\\ n&\ge\sqrt{ \frac{1}{3\times 0.000005}} \approx258.2 \end{align*}1.11.6.31.
Solution.
- When \(0 \leq x \leq1\text{,}\) then \(x^2 \leq 1\) and \(x+1 \geq 1\text{,}\) so \(|f''(x)| = \dfrac{x^2}{|x+1|}\leq \dfrac{1}{1}=1\text{.}\)
-
To find the maximum value of a function over a closed interval, we test the function’s values at the endpoints of the interval and at its critical points inside the interval. The critical points are where the function’s derivative is zero or does not exist.The function we’re trying to maximize is \(|f''(x)| = \frac{x^2}{|x+1|} = \frac{x^2}{x+1}=f''(x)\) (since our interval only contains nonnegative numbers). So, the critical points occur when \(f'''(x) = 0\) or does not exist. We find \(f'''(x)\) Using the quotient rule.\begin{align*} f'''(x)&=\frac{(x+1)(2x)-x^2}{(x+1)^2}=\frac{x^2+2x}{(x+1)^2}\\ 0&=\frac{x(x+2)}{x+1}\\ 0&=x \quad\mbox{or}\quad x=-1\quad\mbox{or}\quad x=-2 \end{align*}The only critical point in \([0,1]\) is \(x=0\text{.}\) So, the extrema of \(f''(x)\) over \([0,1]\) will occur at its endpoints. Indeed, since \(f'''(x) \geq 0\) for all \(x\) in \([0,1]\text{,}\) \(f''(x)\) is increasing over this interval, so its maximum occurs at \(x=1\text{.}\) That is,\begin{equation*} |f''(x)|\leq f''(1)=\frac{1}{2} \end{equation*}
- The absolute error using the midpoint rule is at most \(\dfrac{M(b-a)^3}{24n^2}\text{.}\) Using \(M=1\text{,}\) if we want this to be no more than \(10^{-5}\text{,}\) we find an acceptable value of \(n\) with the following calculation:\begin{align*} \dfrac{M(b-a)^3}{24n^2}&\leq 10^{-5}\\ \dfrac{1}{24n^2}&\leq 10^{-5}&(b-a=1,\,M=1)\\ \frac{10^5}{24} & \leq n^2\\ n & \geq 65 \end{align*}
- The absolute error using the midpoint rule is at most \(\dfrac{M(b-a)^3}{24n^2}\text{.}\) Using \(M=\frac{1}{2}\text{,}\) if we want this to be no more than \(10^{-5}\text{,}\) we find an acceptable value of \(n\) with the following calculation:\begin{align*} \dfrac{M(b-a)^3}{24n^2}&\leq 10^{-5}\\ \dfrac{1}{48n^2}&\leq 10^{-5}&(b-a=1,\,M=\frac{1}{2})\\ \frac{10^5}{48} & \leq n^2\\ n & \geq 46 \end{align*}
1.11.6.32.
Solution.
Because \(x-1 \leq 2\) for every \(x\) in \([1,3]\text{,}\) if \(n^4 \gt \dfrac{24\cdot 2^5}{18}\text{,}\) then \(n^4 \gt \dfrac{24\cdot (x-1)^5}{18}\) for every allowed \(x\text{.}\)
\begin{align*} n^4& \gt \frac{24\cdot 2^5}{18}= \frac{128}{3}\\ n& \gt \sqrt[4]{\frac{128}{3}}\approx 2.6 \end{align*}1.11.6.33.
Solution.
We won’t know the value of the integral exactly, but we’ll have an approximation \(A\) bounded by some positive error bound \(\varepsilon\text{.}\) Then,
\begin{align*} - \varepsilon &\leq \left(\int_1^2 \frac{1}{1+x^2} \dee{x} - A\right) \leq \varepsilon\\ A - \varepsilon &\leq \left(\int_1^2 \frac{1}{1+x^2} \dee{x}\right) \leq A + \varepsilon\\ \text{So, from ($*$), }\qquad \frac{\pi}{4}+A-\varepsilon& \leq \arctan(2) \leq \frac{\pi}{4}+A+\varepsilon \end{align*}
1.12 Improper Integrals
1.12.4 Exercises
1.12.4.1.
Solution.
1.12.4.2.
Solution.
1.12.4.3.
Solution.
1.12.4.4. (✳).
Solution.
1.12.4.5.
Solution.
-
Not enough information to decide. For example, consider \(h(x) = 0\) versus \(h(x) = -1\text{.}\) In both cases, \(h(x) \leq f(x)\text{.}\) However, \(\displaystyle\int_0^\infty 0 \dee{x}\) converges to 0, while \(\displaystyle\int_0^\infty -1 \dee{x}\) diverges.Note: if we had also specified \(0 \leq h(x)\text{,}\) then we would be able to conclude that \(\int_0^\infty h(x) \dee{x}\) converges by the comparison test.
- Not enough information to decide. For example, consider \(h(x)= f(x)\) versus \(h(x) = g(x)\text{.}\) In both cases, \(f(x) \leq h(x) \leq g(x)\text{.}\)
- \(\displaystyle\int_{0\vphantom{\frac12}}^{\infty}h(x) \dee{x}\) converges.
- From the given information, \(|h(x)| \leq 2f(x)\text{.}\)
- We claim \(\displaystyle\int_{0\vphantom{\frac12}}^{\infty} 2f(x) \dee{x} \) converges.
- We can see this by writing \(\displaystyle\int_{0\vphantom{\frac12}}^{\infty} 2f(x) \dee{x}= 2\int_0^{\infty} f(x) \dee{x} \) and noting that the second integral converges.
- Alternately, we can use the limiting comparison test, Theorem 1.12.22. Since \(f(x) \geq 0\text{,}\) \(\displaystyle\int_0^\infty f(x) \dee{x}\) converges, and \(\lim\limits_{x \to \infty}\dfrac{2f(x)}{f(x)}=2\) (the limit exists), we conclude \(\displaystyle\int_0^\infty 2f(x) \dee{x}\) converges.
- So, comparing \(h(x)\) to \(2f(x)\text{,}\) by the comparison test (Theorem 1.12.17) \(\displaystyle\int_0^{\vphantom{\frac12}\infty}h(x) \dee{x}\) converges.
1.12.4.6. (✳).
Solution.
1.12.4.7. (✳).
Solution.
1.12.4.8. (✳).
Solution.
- \(\frac{1}{2x}\) and \(\frac{1}{\sqrt{4x^2-x}}\) are defined and continuous for all \(x \geq 1\text{,}\)
- \(\frac{1}{2x} \geq 0\) for all \(x \geq 1\text{,}\)
- \(\frac{1}{\sqrt{4x^2-x}} \ge\frac{1}{\sqrt{4x^2}} = \frac{1}{2x} \) for all \(x \ge 1\text{,}\) and
- \(\int_1^\infty \frac{1}{2x} \dee{x}\) diverges.
1.12.4.9. (✳).
Solution.
1.12.4.10.
Solution.
These are easy enough to antdifferentiate.
\begin{align*} &=\lim_{a \to \infty }\left[\sin 0 - \sin(- a) \right]+ \lim_{b \to \infty }\left[\sin b - \sin 0 \right]\\ &=\text{DNE} \end{align*}1.12.4.11.
Solution.
1.12.4.12.
Solution.
- Both \(f(x)\) and \(g(x)\) are defined and continuous for all \(x \gt 0\text{,}\) so in particular they are defined and continuous for \(x \geq 10\text{.}\)
- \(g(x) \geq 0\) for all \(x \geq 10\)
- \(\displaystyle\int_{10}^\infty g(x) \dee{x}\) diverges.
- Using l’Hôpital’s rule (5 times!), or simply dividing both the numerator and denominator by \(x^5\) (the common leading term), tells us:\begin{align*} \lim_{x \to \infty} \frac{f(x)}{g(x)}&= \lim_{x \to \infty} \frac{\frac{x^4-5x^3+2x-7}{x^5+3x+8} }{ \frac{1}{x}}= \lim_{x \to \infty} x\cdot\frac{x^4-5x^3+2x-7}{x^5+3x+8}\\ &=\lim_{x \to \infty} \frac{x^5-5x^4+2x^2-7x}{x^5+3x+8} =1 \end{align*}That is, the limit exists and is nonzero.
1.12.4.13.
Solution.
A removable discontinuity doesn’t affect the integral.
\begin{align*} &=\int_0^{10} \frac{1}{x-10} \dee{x}\\ \end{align*}Use the substitution \(u=x-10\text{,}\) \(\dee{u}=\dee{x}\text{.}\) When \(x=0\text{,}\) \(u=-10\text{,}\) and when \(x=10\text{,}\) \(u=0\text{.}\)
\begin{align*} &=\int_{-10}^0 \frac{1}{u} \dee{u} \end{align*}1.12.4.14. (✳).
Solution.
1.12.4.15. (✳).
Solution.
- \(\dfrac{|\sin x|}{x^{3/2}+x^{1/2}}\,, \dfrac{1}{x^{1/2}}\,,\) and \(\dfrac{1}{x^{3/2}}\) are defined and continuous for \(x \gt 0\)
- \(\dfrac{1}{x^{1/2}}\) and \(\dfrac{1}{x^{3/2}}\) are nonnegative for \(x \ge 0\)
- \(\dfrac{|\sin x|}{x^{3/2}+x^{1/2}}\le \dfrac{1}{x^{1/2}}\) for all \(x \gt 0\) and \(\displaystyle\int_0^1\dfrac{1}{x^{1/2}} \dee{x}\) converges, so \(\displaystyle\int_0^{1\vphantom{\frac12}}\dfrac{|\sin x|}{x^{3/2}+x^{1/2}} \dee{x}\) converges.
- \(\dfrac{|\sin x|}{x^{3/2}+x^{1/2}}\le \dfrac{1}{x^{3/2}}\) for all \(x \ge 1\) and \(\displaystyle\int_1^\infty\dfrac{1}{x^{3/2}} \dee{x}\) converges, so \(\displaystyle\int_1^{\infty\vphantom{\frac12}}\dfrac{|\sin x|}{x^{3/2}+x^{1/2}} \dee{x}\) converges.
1.12.4.16. (✳).
Solution.
- \(\frac{x+1}{x^{1/3}(x^2+x+1)}\,, \frac{1}{x^{1/3}}\,,\) and \(\frac{1}{x^{4/3}}\) are defined and continuous for all \(x \gt 0\)
- \(\frac{1}{x^{1/3}}\) and \(\frac{1}{x^{4/3}}\) are positive for all \(x \gt 0\)
- \(\int_0^1 \frac{1}{x^{1/3}} \dee{x}\) and \(\int_1^\infty \frac{1}{x^{4/3}} \dee{x}\) both converge
- \(\lim\limits_{x \to 0}\dfrac{\frac{x+1}{x^{1/3}(x^2+x+1)}}{\frac{1}{x^{1/3}}} =\lim\limits_{x \to 0}\dfrac{x+1}{x^2+x+1}=\dfrac{0+1}{0+0+1}=1 \text{;}\) in particular, this limit exists.
- Using the limiting comparison test (Theorem 1.12.25), \(\int_0^1 \frac{x+1}{x^{1/3}(x^2+x+1)} \dee{x}\) converges.
- \(\lim\limits_{x \to \infty}\dfrac{\frac{x+1}{x^{1/3}(x^2+x+1)}}{\frac{1}{x^{4/3}}} =\lim\limits_{x \to 0}\dfrac{x(x+1)}{x^2+x+1}=1 \text{;}\) in particular, this limit exists.
- Using the limiting comparison test (Theorem 1.12.22), \(\int_1^\infty \frac{x+1}{x^{1/3}(x^2+x+1)} \dee{x}\) converges.
1.12.4.17.
Solution.
1.12.4.18. (✳).
Solution.
1.12.4.19.
Solution.
At this point, we can see that the integral diverges when \(p=1\text{.}\) When \(p \neq 1\text{,}\) we have the limit
\begin{align*} \displaystyle\lim_{t \to \infty}\frac{1/2}{1-p}\left[(t^2+1)^{1-p}-1\right]&=\frac{1/2}{1-p}\left[\lim_{t \to \infty} (t^2+1)^{1-p}\right] - \frac{1/2}{1-p} \end{align*}1.12.4.20.
Solution.
- First, we notice there is only one “source of impropriety:” the domain of integration is infinite. (The integrand has a singularity at \(t=1\text{,}\) but this is not in the domain of integration, so it’s not a problem for us.)
- We should try to get some intuition about whether the integral converges or diverges. When \(t \to \infty\text{,}\) notice the integrand “looks like” the function \(\frac{1}{t^4}\text{.}\) We know \(\int_1^\infty \frac{1}{t^4} \dee{t}\) converges, because it’s a \(p\)-integral with \(p=4 \gt 1\) (see Example 1.12.8). So, our integral probably converges as well. If we were only asked show it converges, we could use a comparison test, but we’re asked more than that.
- Since we guess the integral converges, we’ll need to evaluate it. The integrand is a rational function, and there’s no obvious substitution, so we use partial fractions.
Multiply by the original denominator.
\begin{align*} 1&=(At\!+\!B)(t\!+\!1)(t\!-\!1) + C(t^2\!+\!1)(t\!-\!1) +D(t^2\!+\!1)(t\!+\!1)\tag{$*$}\\ \end{align*}Set \(t=1\text{.}\)
\begin{align*} 1&=0+0+D(2)(2) \qquad \Rightarrow \qquad \color{red}{D=\frac{1}{4}}\\ \end{align*}Set \(t=-1\text{.}\)
\begin{align*} 1&=0+C(2)(-2)+0 \qquad \Rightarrow \qquad \color{red}{C=-\frac{1}{4}}\\ \end{align*}Simplify (\(*\)) using \(D=\frac{1}{4}\) and \(C=-\frac{1}{4}\text{.}\)
\begin{align*} 1&=(At+B)(t+1)(t-1) \textcolor{red}{-\frac{1}{4}}(t^2+1)(t-1) +\textcolor{red}{\frac{1}{4}}(t^2+1)(t+1)\\ &=(At+B)(t+1)(t-1) + \frac{1}{2}(t^2+1)\\ &=At^3+\left(B+\frac{1}{2}\right)t^2-At+\left(\frac{1}{2}-B\right)\\ \end{align*}By matching up coefficients of corresponding powers of \(t\text{,}\) we find \(A=0\) and \(B=-\frac{1}{2}\).
\begin{align*} \int_2^\infty \frac{1}{t^4-1} \dee{t}&=\int_2^\infty \left( \frac{-1/2}{t^2+1} - \frac{1/4}{t+1}+\frac{1/4}{t-1}\right) \dee{t}\\ &=\lim_{R \to \infty}\int_2^R \left( \frac{-1/2}{t^2+1} - \frac{1/4}{t+1}+\frac{1/4}{t-1}\right) \dee{t}\\ &=\lim_{R \to \infty} \left[-\frac{1}{2}\arctan t - \frac{1}{4}\log|t+1|+\frac{1}{4}\log|{t-1}|\right]_2^R\\ &=\lim_{R \to \infty} \left[-\frac{1}{2}\arctan t + \frac{1}{4}\log\left| \frac{t-1}{t+1} \right|\right]_2^R\\ &=\lim_{R \to \infty} \Big(-\frac{1}{2}\arctan R +\frac12\arctan 2 + \frac{1}{4}\log\left| \frac{R-1}{R+1} \right|\\ \amp\hskip0.5in-\frac{1}{4}\log\left| \frac{2-1}{2+1} \right|\Big)\\ \end{align*}We can use l’Hôpital’s rule to see \(\displaystyle\lim_{R \to \infty}\frac{R-1}{R+1}=1\text{.}\) Also note \(-\log (1/3) = \log 3\text{.}\)
\begin{align*} &=-\frac{1}{2}\left(\frac{\pi}{2}\right) +\frac{1}{2}\arctan 2 + \frac{1}{4}\log1+\frac{1}{4}\log3\\ &=\frac{\log 3-\pi}{4} + \frac{1}{2}\arctan 2 \end{align*}1.12.4.21.
Solution.
That looks a lot better. Also, we have a good reason to guess these integrals converge--they look like \(p\)-integrals with \(p=\frac{1}{2}\text{.}\) Let’s take a closer look at each one.
\begin{align*} \int_{-5}^5 \frac{1}{\sqrt {|x|}}\dee{x}&=\int_{-5}^0 \frac{1}{\sqrt {|x|}}\dee{x}+ \int_{0}^5 \frac{1}{\sqrt {|x|}}\dee{x}\\ &=2\int_{0}^5 \frac{1}{\sqrt {|x|}}\dee{x} \qquad\text{(even function)}\\ &=2\int_{0}^5 \frac{1}{\sqrt {x}}\dee{x}\\ \end{align*}This is a \(p\)-integral, with \(p=\frac{1}{2}\text{.}\) By Example 1.12.9 (and Theorem 1.12.20, since the upper limit of integration is not 1), it converges. The other two pieces behave similarly.
\begin{align*} \int_{-5}^5 \frac{1}{\sqrt{|x-1|}}\dee{x}&= \int_{-5}^1 \frac{1}{\sqrt{|x-1|}}\dee{x}+ \int_{1}^5 \frac{1}{\sqrt{|x-1|}}\dee{x}\\ \end{align*}Use \(u=x-1\text{,}\) \(\dee{u}=\dee{x}\)
\begin{align*} &= \int_{-6}^0 \frac{1}{\sqrt{|u|}}\dee{u}+ \int_{0}^4 \frac{1}{\sqrt{|u|}}\dee{x}\\ &= \int_{0}^6 \frac{1}{\sqrt{u}}\dee{u}+ \int_{0}^4 \frac{1}{\sqrt{u}}\dee{x}\\ \end{align*}Since our function is even, we use the reasoning of Example 1.2.10 in the text to consider the area under the curve when \(x \ge 0\text{,}\) rather than when \(x\leq 0\text{.}\) Again, these are \(p\)-integrals with \(p = \frac{1}{2}\text{,}\) so they both converge. Finally:
\begin{align*} \int_{-5}^5\frac{1}{\sqrt{|x-2|}}\dee{x}&= \int_{-5}^2\frac{1}{\sqrt{|x-2|}}\dee{x}+ \int_{2}^5\frac{1}{\sqrt{|x-2|}}\dee{x}\\ \end{align*}Use \(u=x-2\text{,}\) \(\dee{u}=\dee{x}\text{.}\)
\begin{align*} &= \int_{-7}^0\frac{1}{\sqrt{|u|}}\dee{u}+ \int_{0}^3\frac{1}{\sqrt{|u|}}\dee{u}\\ &= \int_{0}^7\frac{1}{\sqrt{u}}\dee{u}+ \int_{0}^3\frac{1}{\sqrt{u}}\dee{u} \end{align*}1.12.4.22.
Solution.
Now let \(u=e^{-x}\text{,}\) \(\dee{v}=\cos x \dee{x}\text{,}\) so \(\dee{u}=-e^{-x} \dee{x}\) and \(v=\sin x\text{.}\)
\begin{align*} &=-e^{-x}\cos x-\left[e^{-x}\sin x + \int e^{-x}\sin x \dee{x} \right]\\ &=-e^{-x}\cos x-e^{-x}\sin x - \int e^{-x}\sin x \dee{x}\\ \end{align*}All together, we found
\begin{align*} \color{red}{\int e^{-x}\sin x \dee{x}} &=-e^{-x}\cos x-e^{-x}\sin x -\color{red}{\int e^{-x}\sin x \dee{x}} +C\\ \color{red}{2\int e^{-x}\sin x \dee{x}} &=-e^{-x}\cos x-e^{-x}\sin x +C\\ \int e^{-x}\sin x \dee{x}&=-\frac{1}{2e^x}(\cos x+\sin x) +C\\ \end{align*}(Remember, since \(C\) is an arbitrary constant, we can rename \(\frac{C}{2}\) to simply \(C\text{.}\)) Now we can evaluate our improper integral.
\begin{align*} \int_0^\infty e^{-x}\sin x \dee{x}&= \lim_{b \to \infty}\int_0^b e^{-x}\sin x \dee{x}\\ &= \lim_{b \to \infty}\left[-\frac{1}{2e^x}(\cos x+\sin x) \right]_0^b\\ &= \lim_{b \to \infty}\left(\frac{1}{2}-\frac{1}{2e^b}(\cos b+\sin b) \right)\\ \end{align*}To find the limit, we use the Squeeze Theorem (see the CLP-1 text). Since \(|\sin b|,|\cos b| \leq 1\) for any \(b\text{,}\) we can use the fact that \(-2 \le \cos b + \sin b \le 2\) for any \(b\text{.}\)
\begin{align*} &\qquad\frac{-2}{2e^b} \leq \frac{1}{2e^b}(\cos b + \sin b) \leq \frac{2}{2e^b}\\ &\qquad\lim_{b \to \infty}\frac{-2}{2e^b} = 0 = \frac{2}{2e^b}\\ &\qquad\mbox{So, }\qquad\lim_{b \to \infty}\left[\frac{1}{2e^b}(\cos b + \sin b)\right]=0\\ \mbox{Therefore, }\qquad\frac{1}{2}&= \lim_{b \to \infty}\left(\frac{1}{2}-\frac{1}{2e^b}(\cos b+\sin b) \right) \end{align*}1.12.4.23. (✳).
Solution.
- \(\frac{\sin^4 x}{x^2}\) and \(\frac{1}{x^2}\) are defined and continuous for every \(x \geq 1\)
- \(0 \leq \frac{\sin^4 x}{x^2} \leq \frac{1^4}{x^2} = \frac{1}{x^2}\) for every \(x \geq 1\)
- \(\int_1^\infty \frac{1}{x^2} \dee{x}\) converges by Example 1.12.8 (it’s a \(p\)-type integral with \(p \gt 1\))
1.12.4.24.
Solution.
-
Solution 1: Let’s try to use a direct comparison. Note \(\dfrac{x}{e^x + \sqrt{x}} \geq 0\) whenever \(x \geq 0\text{.}\) Also note that, for large values of \(x\text{,}\) \(e^x\) is much larger than \(\sqrt{x}\text{.}\) That leads us to consider the following inequalty:\begin{equation*} 0\leq \dfrac{x}{e^x + \sqrt{x}} \leq \dfrac{x}{e^x } \end{equation*}If \(\int_0^\infty \frac{x}{e^x} \dee{x}\) converges, we’re in business. Let’s figure it out. The integrand looks like a candidate for integration by parts: take \(u=x\text{,}\) \(\dee{v} = e^{-x} \dee{x}\text{,}\) so \(\dee{u}=\dee{x}\) and \(v=-e^{-x}\text{.}\)\begin{align*} \int_0^\infty \frac{x}{e^x} \dee{x}&=\lim_{b \to \infty}\int_0^b \frac{x}{e^x} \dee{x}= \lim_{b \to \infty}\left( \left[-\frac{x}{e^x}\right]_0^b + \int_0^b e^{-x} \dee{x} \right)\\ &= \lim_{b \to \infty}\left(-\frac{b}{e^b} +\left[-e^{-x}\right]_0^b \right)= \lim_{b \to \infty}\left(-\frac{b}{e^b} -\frac{1}{e^b}+1 \right)\\ &= \lim_{b \to \infty}\bigg(1-\underbrace{\frac{b+1}{e^b}}_{\atp{\mathrm{num}\to\infty}{\mathrm{den}\to\infty}} \bigg)=\lim_{b \to \infty}\bigg(1-\frac{1}{e^b} \bigg)=1 \end{align*}Using l’Hôpital’s rule, we see \(\int_0^\infty \frac{x}{e^x} \dee{x}\) converges. All together:
- \(\frac{x}{e^x}\) and \(\frac{x}{e^x+\sqrt{x}}\) are defined and continuous for all \(x \geq 0\text{,}\)
- \(\left|\frac{x}{e^x+\sqrt{x}}\right|\leq \frac{x}{e^x}\text{,}\) and
- \(\int_0^\infty \frac{x}{e^x} \dee{x}\) converges.
So, by Theorem 1.12.17, our integral \(\displaystyle\int_0^\infty \frac{x}{e^x+\sqrt{x}} \dee{x}\) converges. -
Solution 2: Let’s try to use a different direct comparison from Solution 1, and avoid integration by parts. We’d like to compare to something like \(\dfrac{1}{e^x}\text{,}\) but the inequality goes the wrong way. So, we make a slight modification: we consider \(2e^{-x/2}\text{.}\) To that end, we claim \(x \lt 2e^{x/2}\) for all \(x \geq 0\text{.}\) We can prove this by noting the following two facts:
- \(0 \lt 2=2e^{0/2}\text{,}\) and
- \(\diff{}{x}\{x\} = 1 \le e^{x/2} = \diff{}{x}\{2e^{x/2}\}\text{.}\)
So, when \(x=0\text{,}\) \(x \lt 2e^{x/2}\text{,}\) and then as \(x\) increases, \(2e^{x/2}\) grows faster than \(x\text{.}\)Now we can make the following comparison:\begin{align*} 0\leq \dfrac{x}{e^x + \sqrt{x}} &\leq \dfrac{x}{e^x } \lt \frac{2e^{x/2}}{e^x} = \frac{2}{e^{x/2}}\\ \end{align*}We have a hunch that \(\int_0^\infty \frac{2}{e^{x/2}} \dee{x}\) converges, just like \(\int_0^\infty \frac{1}{e^{x}} \dee{x}\text{.}\) This is easy enough to prove. We can guess an antiderivative, or use the substitution \(u=x/2\text{.}\)
\begin{align*} \int_0^\infty \frac{2}{e^{x/2}} \dee{x}&=\lim_{R \to \infty} \int_0^R \frac{2}{e^{x/2}} \dee{x} =\lim_{R \to \infty} \left[- \frac{4}{e^{x/2}}\right]_0^R\\ &=\lim_{R \to \infty} \left[\frac{4}{e^{0}}- \frac{4}{e^{R/2}}\right]_0^R=4 \end{align*}Now we know:- \(0 \leq \frac{x}{e^x+\sqrt{x}} \leq \frac{2}{e^{x/2}}\text{,}\) and
- \(\int_0^\infty \frac{2}{e^{x/2}} \dee{x}\) converges.
- Furthermore, \(\frac{x}{e^x+\sqrt{x}}\) and \(\frac{2}{e^{x/2}}\) are defined and continuous for all \(x \geq 0\text{.}\)
By the comparison test (Theorem 1.12.17), we conclude the integral converges. -
Solution 3: Let’s use the limiting comparison test (Theorem 1.12.22). We have a hunch that our integral behaves similarly to \(\int_0^\infty \frac{1}{e^x}\,\dee{x}\text{,}\) which converges (see Example 1.12.18). Unfortunately, if we choose \(g(x) = \frac{1}{e^x}\) (and, of course, \(f(x) = \frac{x}{e^x+\sqrt{x}}\)), then\begin{equation*} \lim_{x \to \infty}\frac{f(x)}{g(x)} = \lim_{x \to \infty}\frac{x}{e^x+\sqrt{x}}\cdot e^x = \lim_{x \to \infty}\frac{x}{1+\underbrace{\tfrac{\sqrt{x}}{e^x}}_{\to 0}} = \infty \end{equation*}That is, the limit does not exist, so the limiting comparison test does not apply. (To find \(\lim\limits_{x \to \infty}\frac{\sqrt{x}}{e^x}\text{,}\) you can use l’Hôpital’s rule.)This setback encourages us to try a slightly different angle. If \(g(x)\) gave larger values, then we could decrease \(\frac{f(x)}{g(x)}\text{.}\) So, let’s try \(g(x) = \frac{1}{e^{x/2}} = e^{-x/2}\text{.}\) Now,\begin{align*} \lim_{x \to \infty}\frac{f(x)}{g(x)} &= \lim_{x \to \infty}\frac{x}{e^x+\sqrt{x}}\div \frac{1}{e^{x/2}}= \lim_{x \to \infty}\frac{x}{e^{x/2}+\frac{\sqrt{x}}{e^{x/2}}} \end{align*}Hmm... this looks hard. Instead of dealing with it directly, let’s use the squeeze theorem (see CLP-1 notes).\begin{gather*} 0 \leq \frac{x}{e^{x/2}+\frac{\sqrt{x}}{e^{x/2 }}}\leq \frac{x}{e^{x/2}} \end{gather*}Using l’Hôpital’s rule,\begin{equation*} \lim_{x \to \infty} \underbrace{\frac{x}{e^{x/2}}}_{\atp{\mathrm{num}\to \infty}{\mathrm{den}\to\infty}} = \lim_{x \to \infty}\frac{1}{\frac{1}{2}e^{x/2}}=0 = \lim_{x \to \infty}0 \end{equation*}So, by the squeeze theorem \(\lim\limits_{x \to 0}\frac{\frac{x}{e^x+\sqrt{x}}}{\frac{1}{e^{x/2}}}=0\text{.}\) Since this limit exists, \(\frac{1}{e^{x/2}}\) is a reasonable function to use in the limiting comparison test (provided its integral converges). So, we need to show that \(\int_0^\infty \frac{1}{e^{x/2}}\,\dee{x}\) converges. This can be done by simply evaluating it:\begin{align*} \int_0^\infty \frac{1}{e^{x/2}}\,\dee{x} \amp= \lim_{b \to \infty}\int_0^b {e^{-x/2}}\,\dee{x} = \lim_{b \to \infty} -\frac{1}{2}\big[e^{-x/2}\big]_0^b\\ \amp= \lim_{b \to \infty} -\frac{1}{2}\left[\frac{1}{e^{b/2}}-1\right] = \frac{1}{2} \end{align*}So, all together:
- The functions \(\frac{x}{e^x+\sqrt{x}}\) and \(\frac{1}{e^{x/2}}\) are defined and continuous for all \(x \geq 0\text{,}\) and \(\frac{1}{e^{x/2}} \ge 0\) for all \(x \ge 0\text{.}\)
- \(\int_0^\infty \frac{1}{e^{x/2}}\,\dee{x}\) converges.
- The limit \(\lim\limits_{x \to \infty}\frac{\frac{x}{e^x}+\sqrt{x}}{\frac{1}{e^{x/2}}}\) exists (it’s equal to 0).
- So, the limiting comparison test (Theorem 1.12.17) tells us that \(\int_0^\infty \frac{x}{e^x+\sqrt{x}}\,\dee{x}\) converges as well.
1.12.4.25. (✳).
Solution.
1.12.4.26.
Solution.
- Since \(f(x)\) is odd, using the reasoning of Example 1.2.11,\begin{align*} \int_{-\infty}^{-1} f(x) \dee{x}\amp=\lim_{t \to \infty}\int_{-t}^{-1}f(x) \dee{x}=\lim_{t \to \infty}-\int_{1}^{t} f(x) \dee{x}\\ \amp=-\lim_{t \to \infty} \int_{1}^{t} f(x) \dee{x} \end{align*}Since \(\displaystyle\int_1^\infty f(x) \dee{x}\) converges, the last limit above converges. Therefore, \(\displaystyle\int_{-\infty\vphantom{\frac12}}^{-1} f(x) \dee{x}\) converges.
- Since \(f(x)\) is even, using the reasoning of Example 1.2.10,\begin{align*} \int_{-\infty}^{-1} f(x) \dee{x}\amp=\lim_{t \to \infty}\int_{-t}^{-1}f(x) \dee{x}=\lim_{t \to \infty}\int_{1}^{t} f(x) \dee{x}\\ \amp=\lim_{t \to \infty} \int_{1}^{t} f(x) \dee{x}\\ \end{align*}and both terms converge, our original integral converges as well.
Since \(\displaystyle\int_1^\infty f(x) \dee{x}\) converges, the last limit above converges. Since \(f(x)\) is continuous everywhere, by Theorem 1.12.20, \(\displaystyle\int_{-1}^\infty f(x) \dee{x}\) converges (note the adjusted lower limit). Then, since
\begin{align*} \int_{-\infty}^\infty f(x) \dee{x}& = \int_{-\infty}^{-1} f(x) \dee{x} + \int_{-1}^\infty f(x) \dee{x} \end{align*}
1.12.4.27.
Solution.
1.13 More Integration Examples
Exercises
1.13.1.
Solution.
1.13.2.
Solution.
- reserve one cosine for the derivative of sine in our substitution, and
- change the rest of the cosines to sines using the identity \(\sin^2x+\cos^2x=1\text{.}\)
1.13.3.
Solution.
1.13.4.
Solution.
1.13.5.
Solution.
- Solution 1: Notice the denominator factors as \((x+1)(3x+1)\text{.}\) Since the integrand is a rational function (the quotient of two polynomials), we can use partial fraction decomposition.\begin{align*} \frac{-2}{3x^2+4x+1}&=\frac{-2}{(x+1)(3x+1)}\\ &=\frac{A}{x+1}+\frac{B}{3x+1}\\ &=\frac{A(3x+1)+B(x+1)}{(x+1)(3x+1)}\\ &=\frac{(3A+B)x+(A+B)}{(x+1)(3x+1)}\\ \end{align*}
So:
\begin{align*} -2&=(3A+B)x+(A+B)\\ 0&=3A+B \mbox{ and }-2=A+B\\ B&=-3A \mbox{ and hence } -2=A+(-3A)\\ {\color{red}{A}}&{\textcolor{red}{=1}} \mbox{ so then } {\color{red}{B=-3}}\\ \end{align*}So now:
\begin{align*} \frac{-2}{3x^2+4x+1}&=\frac{1}{x+1}-\frac{3}{3x+1}\\ \int \frac{-2}{3x^2+4x+1}\dee{x}&=\int\left(\frac{1}{x+1}-\frac{3}{3x+1}\right)\dee{x}\\ &=\log|x+1|-\log|3x+1|+C\\ &=\log\left|\dfrac{x+1}{3x+1}\right|+C \end{align*} -
Solution 2: The previous solution is probably the nicest. However, for the foolhardy or the brave, this integral can also be evaluated using trigonometric substitution.We start by completing the square on the denominator.\begin{align*} 3x^2+4x+1 &= 3\left(x^2+\frac{4}{3}x+\frac{1}{3}\right)\\ &=3\left(x^2+2\cdot\frac{2}{3}x+\frac{4}{9}-\frac{4}{9}+\frac{1}{3}\right)\\ &=3\left(\left(x+\frac{2}{3}\right)^2-\frac{4}{9}+\frac{3}{9}\right)\\ &=3\left(\left(x+\frac{2}{3}\right)^2-\frac{1}{9}\right)\\ &=3\left(x+\frac{2}{3}\right)^2-\frac{1}{3}\\ \end{align*}
This has the form of a function minus a constant, which matches the trigonometric identity \(\sec^2\theta - 1 = \tan^2\theta\text{.}\) Multiplying through by \(\frac{1}{3}\text{,}\) we see we can use the identity \(\frac{1}{3}\sec^2\theta - \frac{1}{3} = \frac{1}{3}\tan^2\theta\text{.}\) So, to get the substitution right, we want to choose a substitution that makes the following true:
\begin{align*} 3\left(x+\frac{2}{3}\right)^2-\frac{1}{3}&=\frac{1}{3}\sec^2\theta-\frac{1}{3}\\ 3\left(x+\frac{2}{3}\right)^2&=\frac{1}{3}\sec^2\theta\\ 9\left(x+\frac{2}{3}\right)^2&=\sec^2\theta\\ \color{red}{3x+2}&=\color{blue}{\sec\theta}\\ \end{align*}And, accordingly:
\begin{align*} \color{red}{3\dee{x}}&=\color{blue}{\sec\theta\tan\theta \dee{\theta}} \end{align*}Now, let’s simplify a little and use this substitution on our integral:\begin{align*} \amp\int\dfrac{-2}{3x^2+4x+1}\dee{x} =\int\dfrac{-2}{3\left(x+\frac{2}{3}\right)^2-\frac{1}{3}}\dee{x}\\ &\hskip0.5in=\int\dfrac{-2}{9\left(x+\frac{2}{3}\right)^2-1}3\dee{x}\\ &\hskip0.5in=\int\dfrac{-2}{\left({\color{red}{3x+2}}\right)^2-1}\textcolor{red}{3\dee{x}}\\ &\hskip0.5in=\int\dfrac{-2}{\left({\color{blue}{\sec\theta}}\right)^2-1}\color{blue}{\sec\theta\tan\theta \dee{\theta}}\\ &\hskip0.5in=\int\dfrac{-2}{\tan^2\theta}\sec\theta\tan\theta \dee{\theta}\\ &=\hskip0.5in\int -2\dfrac{\sec\theta}{\tan\theta}\dee{\theta}\\ &\hskip0.5in=\int -2\dfrac{1}{\cos\theta}\cdot\dfrac{\cos\theta}{\sin\theta} \dee{\theta}\\ &\hskip0.5in=\int -2\dfrac{1}{\sin\theta} \dee{\theta}\\ &\hskip0.5in=\int -2\csc\theta \dee{\theta}\\ \end{align*}Using the result of Example 1.8.21, or a table of integrals:
\begin{align*} &=2\log\left|\csc\theta+\cot\theta \right|+C \end{align*}Our final task is to translate this back from \(\theta\) to \(x\text{.}\) Recall we used the substitution \(3x+2=\sec\theta\text{.}\) Using this information, and \(\sec\theta = \dfrac{\mathrm{hypotenuse}}{\mathrm{adjacent}}\text{,}\) we can fill in two sides of a right triangle with angle \(\theta\text{.}\) The Pythagorean theorem tells us the third side (opposite to \(\theta\)) has measure \(\sqrt{(3x+2)^2-1}=\sqrt{9x^2-12x+3}\text{.}\)\begin{align*} \amp 2\log\left|\csc\theta+\cot\theta \right|+C\\ \hskip0.5in&= 2\log\left|\dfrac{3x+2}{\sqrt{9x^2+12x+3}}+\dfrac{1}{\sqrt{9x^2+12x+3}} \right|+C\\ &\hskip0.5in= 2\log\left|\dfrac{3x+3}{\sqrt{9x^2+12x+3}}\right|+C\\ &\hskip0.5in= \log\left|\dfrac{(3x+3)^2}{\sqrt{9x^2+12x+3}^2}\right|+C\\ &\hskip0.5in= \log\left|\dfrac{(3x+3)^2}{9x^2+12x+3}\right|+C\\ &\hskip0.5in= \log\left|\dfrac{9(x+1)^2}{3(3x+1)(x+1)}\right|+C\\ &\hskip0.5in= \log\left|\dfrac{3(x+1)^2}{(3x+1)(x+1)}\right|+C\\ &\hskip0.5in= \log\left|\dfrac{3(x+1)}{3x+1}\right|+C\\ &\hskip0.5in= \log\left|\dfrac{x+1}{3x+1}\right|+\log3+C\\ \end{align*}Since \(C\) is an arbitrary constant, we can write our final answer as
\begin{align*} & \log\left|\dfrac{x+1}{3x+1}\right|+C \end{align*}
1.13.6.
Solution.
We use the Fundamental Theorem of Calculus Part 2 to evaluate the definite integral.
\begin{align*} \int_1^2 x^3\log x \dee{x}&= \left[\frac{1}{3}x^3\log x - \frac{1}{9}x^3\right]_1^2\\ &=\left[\frac{1}{3}2^3\log 2 - \frac{1}{9}2^3 \right] - \left[\frac{1}{3}1^3\log 1 - \frac{1}{9}1^3 \right]\\ &=\frac{8\log 2}{3}-\frac{8}{9}-0+\frac{1}{9}\\ &=\frac{8}{3}\log2-\frac{7}{9} \end{align*}1.13.7. (✳).
Solution.
1.13.8. (✳).
Solution.
1.13.9. (✳).
Solution.
1.13.10. (✳).
Solution.
1.13.11. (✳).
Solution.
1.13.12.
Solution.
Setting \(x=3\) and \(x=2\) gives us
\begin{align*} \color{red}{B}&\color{red}{=12,\quad A=-12} \end{align*}1.13.13. (✳).
Solution.
Multiply by the original denominator.
\begin{align*} 1&=a(x^2+1)+(bx+c)(x-1)\tag{$*$} \end{align*}To antidifferentiate the second piece, we split it into two integrals: one that can be handled with the substitution \(u=x^2+1\text{,}\) and another that looks like the derivative of arctangent.
\begin{align*} &=\int\Big(\frac{1/2}{x-1}-\frac{x/2}{x^2+1} -\frac{1/2}{x^2+1}\Big)\,\dee{x}\cr &=\int\Big(\frac{1/2}{x-1}-\frac{1}{4}\cdot\frac{2x}{x^2+1} -\frac{1/2}{x^2+1}\Big)\,\dee{x}\cr &=\frac{1}{2}\log|x-1|-\frac{1}{4}\log(x^2+1)-\frac{1}{2}\arctan x+C \end{align*}1.13.14. (✳).
Solution.
1.13.15. (✳).
Solution.
- Solution 1: Integrate by parts, using \(u=\log(1+x^2)\) and \(\dee{v}=x\,\dee{x}\text{,}\) so that \(\dee{u}=\frac{2x}{1+x^2}\text{,}\) \(v=\frac{x^2}{2}\text{.}\)\begin{align*} \int_0^1 x\log(1+x^2)\,\dee{x} &=\Big[\frac{1}{2} x^2\log(1+x^2)\Big]_0^1-\int_0^1 \frac{x^3}{1+x^2}\,\dee{y}\\ \amp=\frac{1}{2} \log 2 -\int_0^1\Big[x-\frac{x}{1+x^2}\Big]\,\dee{x}\\ &=\frac{1}{2} \log 2 -\Big[\frac{x^2}{2}-\frac{1}{2}\log(1+x^2)\Big]_0^1\\ \amp=\log 2-\frac{1}{2}\approx 0.193 \end{align*}
- Solution 2: First substitute \(y=1+x^2\text{,}\) \(\dee{y}=2x\,\dee{x}\text{.}\)\begin{gather*} \int_0^1 x\log(1+x^2)\,\dee{x} =\frac{1}{2}\int_1^2\log y\ \dee{y} \end{gather*}Then integrate by parts, using \(u=\log y\) and \(\dee{v}=\dee{y}\text{,}\) so that \(\dee{u}=\frac{1}{y}\text{,}\) \(v=y\text{.}\)\begin{align*} \int_0^1 x\log(1+x^2)\,\dee{x} \amp=\frac{1}{2}\int_1^2\log y\ \dee{y}\\ \amp=\Big[\frac{1}{2} y\log y\Big]_1^2-\frac{1}{2}\int_1^2 y\frac{1}{y}\,\dee{y} =\log 2-\frac{1}{2}\\ \amp\approx 0.193 \end{align*}
1.13.16. (✳).
Solution.
1.13.17. (✳).
Solution.
1.13.18. (✳).
Solution.
1.13.19.
Solution.
1.13.20. (✳).
Solution.
1.13.21. (✳).
Solution.
1.13.22.
Solution.
- Solution 1: Let’s use the substitution \(u=x-1\text{,}\) \(\dee{u}=\dee{x}\text{.}\)\begin{align*} \int x\sqrt{x-1}\dee{x}&=\int(u+1)\sqrt{u}\dee{u}\\ &=\int \left(u^{3/2}+u^{1/2}\right)\dee{u}\\ &=\frac{2}{5}u^{5/2} + \frac{2}{3}u^{3/2}+C\\ &=\frac{2}{5}(x-1)^{5/2} + \frac{2}{3}(x-1)^{3/2}+C \end{align*}
- Solution 2: We have an integrand with \(x\) multiplied by something integrable. So, if we use integration by parts with \(u=x\) and \(\dee{v} = \sqrt{x-1}\dee{x}\text{,}\) then \(\dee{u}=\dee{x}\) (that is, the \(x\) goes away) and \(v = \frac{2}{3}(x-1)^{3/2}\text{.}\)\begin{align*} \int x\sqrt{x-1}\dee{x}&=\frac{2}{3}x\sqrt{x-1}^3 - \frac{2}{3}\int(x-1)^{3/2}\dee{x}\\ &= \frac{2}{3}x\sqrt{x-1}^3 - \frac{2}{3}\left(\frac{2}{5}(x-1)^{5/2}\right)+C\\ &=\frac{2}{3}\sqrt{x-1}\left(x(x-1)-\frac{2}{5}(x-1)^2\right)+C\\ &=\frac{2}{15}\sqrt{x-1}\cdot(3x^2-x-2)+C\\ &=\frac{2}{15}\sqrt{x-1}\cdot(3(x^2-2x+1)+5x-5)+C\\ &=\frac{2}{15}\sqrt{x-1}\cdot(3(x-1)^2+5(x-1))+C\\ &=\frac{2}{15}\cdot 3\sqrt{x-1}^5 +\frac{2}{15}\cdot 5\sqrt{x-1}^3+C\\ &=\frac{2}{5}\sqrt{x-1}^5 +\frac{2}{3}\sqrt{x-1}^3+C \end{align*}
1.13.23.
Solution.
1.13.24.
Solution.
1.13.25.
Solution.
So, by matching coefficients:
\begin{align*} &A=3, 2A+B=4,\mbox{ and } A+B+C=6\\ &\color{red}{A=3,\quad B=-2,\quad C=5}\\ \end{align*}Therefore:
\begin{align*} \frac{3x^2+4x+6}{(x+1)^3} &= \frac{\textcolor{red}{3}}{x+1}+\frac{\textcolor{red}{-2}}{(x+1)^2}+\frac{\textcolor{red}{5}}{(x+1)^3}\\ \end{align*}Now, the integration is easy, with a substitution of \(u=x+1\) and \(\dee{u}=\dee{x}\text{:}\)
\begin{align*} \int\frac{3x^2+4x+6}{(x+1)^3} \dee{x}&= \int\left(\frac{3}{x+1}+\frac{-2}{(x+1)^2}+\frac{5}{(x+1)^3}\right)\dee{x}\\ &=\int \left( 3u^{-1}-2u^{-2}+5u^{-3}\right)\dee{u}\\ &=3\log|u|+2u^{-1}-\frac{5}{2} u^{-2}+C\\ &=3\log|x+1|+\frac{2}{x+1}-\frac{5}{2(x+1)^2}+C \end{align*}1.13.26.
Solution.
First: complete the square
\begin{align*} \int\frac{1}{x^2+x+1}\dee{x}&= \int\frac{1}{x^2+x+\frac{1}{4}+\frac{3}{4}}\dee{x} = \int\frac{1}{\left(x+\frac{1}{2}\right)^2+\frac{3}{4}}\dee{x}\\ \end{align*}Second: get the denominator in the form \(u^2+1\text{.}\) To do this, we need to fix the constant
\begin{align*} &=\int\left(\frac{1}{\left(x+\frac{1}{2}\right)^2+\frac{3}{4}}\right)\left(\frac{\frac{4}{3}}{\frac{4}{3}}\right)\dee{x}\\ &=\frac{4}{3}\int\frac{1}{\frac{4}{3}\cdot\left(x+\frac{1}{2}\right)^2+1}\dee{x}\\ \end{align*}Now a quick wiggle to make that first part of the denominator into something squared again:
\begin{align*} &=\frac{4}{3}\int\frac{1}{\left(\frac{2}{\sqrt{3}}x+\frac{1}{\sqrt{3}}\right)^2+1}\dee{x}\\ \end{align*}Now we see that \(u=\dfrac{2}{\sqrt{3}}x+\dfrac{1}{\sqrt{3}}\text{,}\) \(\dee{u}=\dfrac{2}{\sqrt3}\dee{x}\) will do the job
\begin{align*} &=\frac{4}{3}\int\frac{1}{u^2+1}\cdot\frac{\sqrt{3}}{2}\dee{u} =\frac{2}{\sqrt3}\int\frac{1}{u^2+1}\dee{u}\\ &=\frac{2}{\sqrt3}\arctan u +C\\ &=\frac{2}{\sqrt3}\arctan\left(\frac{2}{\sqrt3}x+\frac{1}{\sqrt3}\right) +C \end{align*}1.13.27.
Solution.
1.13.28.
Solution.
When \(x=-1\text{,}\) we see \(1=3A\text{,}\) so \(\textcolor{red}{\frac{1}{3}=A}\text{.}\) We plug this into (\(*\)).
\begin{align*} 1&=\textcolor{red}{\frac{1}{3}}(x^2-x+1)+(Bx+C)(x+1)\\ -\frac{1}{3}x^2+\frac{1}{3}x+\frac{2}{3}&=Bx^2+(B+C)x+C\\ \end{align*}Matching up coefficients of corresponding power of \(x\text{,}\) we see \(\textcolor{red}{B = -\frac{1}{3}}\) and \(\textcolor{red}{C = \frac{2}{3}}\text{.}\)
\begin{align*} \int \frac{1}{x^3+1}\dee{x}&=\int \left(\frac{\textcolor{red}{1/3}}{x+1} \textcolor{red}{-} \frac{\textcolor{red}{\frac{1}{3}}x-\textcolor{red}{\frac{2}{3}}}{x^2-x+1}\right)\dee{x}\\ \end{align*}To integrate the second fraction, we break it up into two pieces: one we can integrate using the substitution \(u=x^2-x+1\text{,}\) the other will look like the derivative of arctangent.
\begin{align*} &=\frac{1}{3}\log|x+1| -\int\frac{\frac{1}{3}x-\frac{1}{6}-\frac{1}{2}}{x^2-x+1}\dee{x}\\ &=\frac{1}{3}\log|x+1| - \frac{1}{6}\int\frac{2x-1}{x^2-x+1}\dee{x} + \frac{1}{2}\int \frac{1}{(x-\frac{1}{2})^2+\frac{3}{4}}\dee{x}\\ &=\frac{1}{3}\log|x+1| - \frac{1}{6}\log|x^2-x+1| + \frac{1}{2}\int\frac{1}{\frac{3}{4}\left(\left(\frac{2x-1}{\sqrt{3}}\right)^2+1\right)} \dee{x}\\ &=\frac{1}{3}\log|x+1| - \frac{1}{6}\log|x^2-x+1| + \frac{2}{3}\int\frac{1}{\left(\frac{2x-1}{\sqrt{3}}\right)^2+1} \dee{x}\\ \end{align*}Let \(u= \frac{2x-1}{\sqrt3}\text{,}\) \(\dee{u} = \frac{2}{\sqrt 3}\dee{x}\text{.}\)
\begin{align*} &=\frac{1}{3}\log|x+1| - \frac{1}{6}\log|x^2-x+1| + \frac{1}{\sqrt3}\int\frac{1}{u^2+1} \dee{x}\\ &=\frac{1}{3}\log|x+1| - \frac{1}{6}\log|x^2-x+1| + \frac{1}{\sqrt3}\arctan\left(\frac{2x-1}{\sqrt3}\right)+C \end{align*}1.13.29.
Solution.
- Option 1: Let \(u=1-x^2\text{.}\) Then \(-\frac{1}{2}\dee{u}=\dee{x}\text{,}\) and \(x^2 = 1-u\text{.}\)\begin{align*} \int (3x)^2\arcsin x \dee{x}&= 3x^3\arcsin x -\int \frac{3x^3}{\sqrt{1-x^2}}\dee{x}\\ &=3x^3 \arcsin x - 3\int\frac{x^2}{\sqrt{1-x^2}}\cdot x\dee{x}\\ &=3x^3 \arcsin x + \frac{3}{2}\int\frac{1-u}{\sqrt{u}}\dee{u}\\ &=3x^3 \arcsin x + \frac{3}{2}\int\left(u^{-1/2} - u^{1/2}\right)\dee{u}\\ &=3x^3 \arcsin x + \frac{3}{2}\left(2u^{1/2} - \frac{2}{3}u^{3/2}\right)+C\\ &=3x^3 \arcsin x + 3\sqrt{1-x^2} - \sqrt{1-x^2}^3+C \end{align*}
-
Option 2: If we let \(x=\sin\theta\text{,}\) then \(\sqrt{1-x^2}=\sqrt{\cos^2\theta}=\cos\theta\text{.}\) So let’s use the substitution \(x=\sin\theta\text{,}\) \(\dee{x}=\cos\theta \dee{\theta}\text{.}\)\begin{align*} \int (3x)^2\arcsin x \dee{x}&= 3x^3\arcsin x -\int \frac{3x^3}{\sqrt{1-x^2}}\dee{x}\\ &=3x^3\arcsin x - \int \frac{3\sin^3\theta}{\sqrt{1-\sin^2\theta}}\cos\theta \dee{\theta}\\ &= 3x^3\arcsin x - \int {3\sin^3\theta \dee{\theta}} \end{align*}\begin{align*} \amp 3x^3\arcsin x - \int {3\sin^3\theta \dee{\theta}}=3x^3\arcsin x - 3\int\sin^2\theta \sin\theta \dee{\theta}\\ &\hskip0.5in=3x^3\arcsin x - 3\int(1-\cos^2\theta) \sin\theta \dee{\theta}\\ &\hskip0.5in=3x^3\arcsin x + 3\int(1-u^2)\dee{u}\\ &\hskip0.5in=3x^3\arcsin x + 3 \left(u-\frac{1}{3}u^3\right) + C\\ &\hskip0.5in=3x^3\arcsin x + 3u-u^3 + C\\ &\hskip0.5in=3x^3\arcsin x + 3\cos\theta-\cos^3\theta + C \end{align*}Recall \(x=\sin\theta\text{;}\) so we draw a triangle with angle \(\theta\text{,}\) opposite side \(x\text{,}\) hypotenuse 1. Then by Pythagoras, adjacent side is \(\sqrt{1-x^2}\text{,}\) so \(\cos\theta = \sqrt{1-x^2}\text{.}\)\begin{align*} \amp\int (3x)^2\arcsin x \dee{x}\\ &\hskip0.5in=3x^3\arcsin x + 3\sqrt{1-x^2}-(1-x^2)^{3/2} + C \end{align*}
1.13.30.
Solution.
Over the interval \([0,\frac{\pi}{2}]\text{,}\) \(\cos (t/2) \gt 0\text{,}\) so we can drop the absolute values.
\begin{align*} &= \sqrt{2}\int_0^{\pi/2} \cos (t/2)\dee{t} = \sqrt{2}\left[2\sin\left(\frac{t}{2}\right)\right]_0^{\pi/2}\\ &=2\sqrt{2}\sin\left(\frac{\pi}{4}\right) = {2} \end{align*}1.13.31.
Solution.
- Solution 1: Using logarithm rules, \(\log\sqrt{x} = \log\left(x^{1/2}\right)=\frac{1}{2}\log x\text{,}\) so we can simplify:\begin{equation*} \int_1^e \frac{\log\sqrt{x}}{{x}}\dee{x} =\int_1^e \frac{\log{x}}{{2x}}\dee{x} \end{equation*}We use the substitution \(u=\log x\text{,}\) \(\dee{u}=\frac{1}{x}\dee{x}\text{:}\)\begin{align*} \int_1^e \frac{\log{x}}{{2x}}\dee{x}&=\frac{1}{2}\int_1^e \underbrace{\log(x)}_{u}\cdot\underbrace{\frac{1}{x}\,\dee{x}}_{\dee{u}}\\ &=\frac{1}{2}\int_{\log(1)}^{\log(e)}u\,\dee{u}\\ &=\frac{1}{2}\int_{0}^{1}u\,\dee{u}\\ &=\frac{1}{2}\left[ \frac{1}{2}u^2\right]_0^1\\ &=\frac{1}{2}\left[\frac{1}{2}-0\right]=\dfrac{1}{4} \end{align*}
- Solution 2: We use the substitution \(u=\log\sqrt{x}\text{.}\) Then \(\dfrac{\dee{u}}{\dee{x}}=\dfrac{1}{\sqrt{x}}\cdot\dfrac{1}{2\sqrt{x}}=\dfrac{1}{2x}\text{,}\) hence \(2\dee{u}=\dfrac{1}{x}\dee{x}\text{.}\) This fits our integral nicely!\begin{align*} \int_1^e\frac{\log\sqrt{x}}{x}\dee{x} &= \int_{\log\sqrt{1}}^{\log\sqrt{e}}u\cdot 2\dee{u}\\ &=\Big[u^2\Big]_{0}^{1/2}\\ &=\left(\frac{1}{2}\right)^2-0^2=\frac{1}{4} \end{align*}
1.13.32.
Solution.
1.13.33. (✳).
Solution.
1.13.34. (✳).
Solution.
1.13.35.
Solution.
Now we’re back in familiar territory. Let \(u=\sin \theta\text{,}\) \(\dee{u}=\cos \theta\dee{\theta}\text{.}\)
\begin{align*} &=-2\int\sqrt{1-\sin^2 \theta }\cos\theta\dee{\theta}\\ &=-2\int\cos^2\theta\dee{\theta}\\ &=-\int \big(1+\cos(2\theta)\big)\dee{\theta}\\ &=-\theta - \frac{1}{2}\sin(2\theta)+C\\ &=-\theta - \sin\theta\cos\theta+C\\ &=-\arcsin u - u\sqrt{1-u^2}+C \tag{$*$}\\ &=-\arcsin (\sqrt{1-x})-\sqrt{1-x}\sqrt{x}+C \end{align*}1.13.36.
Solution.
This is more familiar. We use integration by parts with \(\dee{v} = e^u\dee{u}\text{,}\) \(v=e^u\text{.}\) Conveniently, the “\(u\)” we brought in with the substitution is what we want to use for the “\(u\)” in integration by parts, so we don’t have to change the names of our variables.
\begin{align*} &=\big[ue^u\big]_1^e - \int_1^e e^u\dee{u}\\ &=e\cdot e^e-e-e^e+e=e^e(e-1) \end{align*}1.13.37.
Solution.
1.13.38.
Solution.
Now we use integration by parts.
\begin{align*} \int \frac{x\sin x}{\cos^2 x}\dee{x}&=x\sec x - \int \sec x \dee{x}= x\sec x - \log|\sec x + \tan x|+C \end{align*}1.13.39.
Solution.
Now, if \(n\neq -1\) and \(n \neq -2\text{,}\) we can just use the power rule:
\begin{align*} &=\frac{u^{(n+2)}}{n+2}-a\frac{u^{n+1}}{n+1}+C\\ &=\frac{(x+a)^{(n+2)}}{n+2}-a\frac{(x+a)^{n+1}}{n+1}+C\\ \end{align*}If \(n=-1\text{,}\) then
\begin{align*} \int x(x+a)^n\dee{x}&=\int \big(u^{n+1}-au^n\big)\dee{u} =\int \left(1 - \frac{a}{u}\right)\dee{u}\\ &= u - a\log|u|+C = (x+a)-a\log|x+a|+C\\ \end{align*}If \(n=-2\text{,}\) then
\begin{align*} \int x(x+a)^n\dee{x}&=\int \big(u^{n+1}-au^n\big)\dee{u}=\int\left( \frac{1}{u} - au^{-2}\right)\dee{u}\\ & = \log|u| + \frac{a}{u}+C= \log|x+a| + \frac{a}{x+a}+C\\ \end{align*}All together,
\begin{align*} \int x(x+a)^n\dee{x}&=\begin{cases} \frac{(x+a)^{(n+2)}}{n+2}-a\frac{(x+a)^{n+1}}{n+1}+C&\text{ if } n \neq -1,-2\\ (x+a)-a\log|a+x|+C & \text{ if } n=-1\\ \log|x+a| +\frac{a}{x+a}+C& \text{ if } n=-2 \end{cases} \end{align*}1.13.40.
Solution.
Since the coefficient of \(x^4\) on the left-hand is 1, we may assume \(a=d=1\text{.}\)
\begin{align*} x^4+1&=(x^2+bx+c)(x^2+ex+f)\\ \end{align*}Since the constant term is 1, \(cf=1\text{.}\) That is, \(f = \frac{1}{c}\text{.}\)
\begin{align*} x^4+1&=(x^2+bx+c)(x^2+ex+1/c)\\ &=x^4 + \underbrace{\vphantom{\bigg(}(b+e)}_{(1)}x^3+\underbrace{\left(\frac{1}{c}+be+c\right)}_{(3)}x^2+\underbrace{\left(\frac{b}{c}+ec\right)}_{(2)}x+1 \end{align*}- The coefficient of \(x^3\) tells us \(e=-b\text{.}\)
- Then the coefficient of \(x\) tells us \(0=\frac{b}{c}+ec = \frac{b}{c}-bc\text{.}\) So, \(c = \frac{1}{c}\text{,}\) hence \(c=\pm 1\text{.}\)
- Finally, the coefficient of \(x^2\) tells us \(0=\frac{1}{c}+be + c = \frac{1}{c}-b^2+c\text{.}\) Since \(-b^2\) is negative (or zero), \(\frac{1}{c}+c\) is positive, so \(c=1\text{.}\) That is, \(0=1-b^2+1\text{.}\) So, \(b = \sqrt{2}\text{.}\)
From the coefficient of \(x^3\text{,}\) we see \(C=-A\text{.}\)
\begin{align*} 2x^2&= (B+D-2\sqrt{2}A)x^2+(-\sqrt2B+\sqrt2D)x+(B+D)\\ \end{align*}From the constant term, we see \(D=-B\text{.}\)
\begin{align*} 2x^2&= (-2\sqrt{2}A)x^2+(-2\sqrt2B)x \end{align*}To integrate, we want to break the fractions into two pieces each: one we can integrate with a substitution \(u=x^2\pm\sqrt2x+1\) , \(\dee{u} =\big( 2x\pm \sqrt{2}\big)\dee{x}\) (shown in blue), and one that looks like the derivative of arctangent (shown in red).
\begin{align*} &=\frac{1}{\sqrt 2}\int\left(\frac{\textcolor{blue}{- x -\frac{\sqrt 2}{2}} +\textcolor{red}{\frac{\sqrt2}{2}}}{x^2+\sqrt2x+1}+\frac{\textcolor{blue}{x-\frac{\sqrt2}{2}}+\textcolor{red}{\frac{\sqrt2}{2}}}{x^2-\sqrt2x+1}\right)\dee{x}\\ &=\frac{1}{\sqrt 2} \int\bigg(\textcolor{blue}{\frac{-\frac{1}{2}(2 x +{\sqrt 2})}{x^2+\sqrt2x+1}} +\textcolor{red}{\frac{\frac{\sqrt2}{2}}{x^2+\sqrt2x+1}}\\ \amp\hskip1in +\textcolor{blue}{\frac{\frac{1}{2}(2x-\sqrt2)}{x^2-\sqrt2x+1}}+\textcolor{red}{\frac{\frac{\sqrt2}{2}}{x^2-\sqrt2x+1}}\bigg)\dee{x}\\ &=\frac{1}{\sqrt 2}\Bigg(\textcolor{blue}{ -\frac{1}{2}\log\left| x^2+\sqrt2x+1\right|} +\textcolor{red}{\int\frac{\frac{\sqrt2}{2}}{x^2+\sqrt2x+1}\dee{x}}\\ &\qquad+ \textcolor{blue}{\frac{1}{2}\log|x^2-\sqrt2x+1|} +\textcolor{red}{ \int \frac{\frac{\sqrt2}{2}}{x^2-\sqrt2x+1}\dee{x}} \Bigg)\\ \end{align*}We use logarithm rules to compress our work. In order to evaluate the remaining integrals, we complete the squares of the denominators.
\begin{align*} &=\frac{1}{\sqrt 2}\Bigg(\textcolor{blue}{ \frac{1}{2}\log\left|\tfrac{x^2-\sqrt2x+1}{ x^2+\sqrt2x+1}\right|} +\textcolor{red}{\int\frac{\frac{\sqrt2}{2}}{\left(x+\frac{1}{\sqrt2}\right)^2+\frac{1}{2}}\dee{x}}\\ \amp\hskip1in + \textcolor{red}{\int \frac{\frac{\sqrt2}{2}}{\left(x-\frac{1}{\sqrt2}\right)^2+\frac{1}{2}}\dee{x}} \Bigg)\\ &=\frac{1}{\sqrt 2}\Bigg( \textcolor{blue}{\frac{1}{2}\log\left|\tfrac{x^2-\sqrt2x+1}{ x^2+\sqrt2x+1}\right|} +\textcolor{red}{\int\frac{{\sqrt2}}{\left(\sqrt{2}x+1\right)^2+1}\dee{x}}\\ \amp\hskip1in+ \textcolor{red}{ \int \frac{{\sqrt2}}{\left(\sqrt{2}x-1\right)^2+1}\dee{x}} \Bigg)\\ \end{align*}Now, we can either guess the antiderivatives of the remaining integrals, or use the substitutions \(u=(\sqrt{2}x\pm1)\text{.}\)
\begin{align*} &=\frac{1}{\sqrt 2}\Bigg( \textcolor{blue}{\frac{1}{2}\log\left|\tfrac{x^2-\sqrt2x+1}{ x^2+\sqrt2x+1}\right|} +\textcolor{red}{\arctan\left(\sqrt{2}x\!+\!1\right)}+\textcolor{red}{ \arctan\left(\sqrt{2}x\!-\!1\right)} \Bigg)\!+\!C \end{align*}
2 Applications of Integration
2.1 Work
2.1.2 Exercises
2.1.2.1.
Solution.
2.1.2.2.
Solution.
2.1.2.3.
Solution.
-
We defined \(\Delta x = \frac{b-a}{n}\text{:}\) that is, the length of one interval, when we chop \([a,b]\) into \(n\) of them. If \(b\) and \(a\) are measured in metres, then \(\Delta x\) is measured in metres as well. So, the units of \(\Delta x\) are metres.Put another way, since \(a\) and \(b\) both describe a quantity in metres, \(b-a\) describes a quantity in metres as well. (When we add or subtract quantities of the same units, their sum or difference is given in the same units.) Since \(n\) is a unitless quantity (simply a number: not “\(n\) kg” or “\(n\) m”), \(\frac{b-a}{n}\) still describes a quantity in metres. (If I have 6 metres of cloth, and I cut it into 3 pieces, each piece has \(\frac{6}{3}=2\) metres — not 2 kilograms, or 2 metres per second.)
- Since \(F(x)\) is measured in kilogram-metres per second squared (newtons), the units of \(F(x_i)\) are kilogram-metres per second squared (newtons).
- \(W\) is calculated by adding up summands of the form \(F(x_i)\Delta x\text{.}\) The units of \(F(x_i)\Delta x\) are the products of the units of \(F(x_i)\) with the units of \(\Delta x\text{.}\) That is, the units of \(F(x_i)\Delta x\) are \(\left(\frac{\text{kg}\cdot\text{m}}{\text{sec}^2}\right)\left(\text{m}\right) = \frac{\text{kg}\cdot\text{m}^2}{\text{sec}^2} = J\text{.}\) The sum of terms given in joules is itself given in joules, so the units of \(W\) are joules.
2.1.2.4.
Solution.
2.1.2.5.
Solution.
- Solution 1: Since the force required to stretch the spring is proportional to the amount stretched, and the force acting on the spring is proportional to the mass hanging from it, we conclude the amount the spring stretches is proportional to the mass hung from it. So, if 1 kg stretches it 1 cm, then 10 kg will stretch it 10 cm. We should mark the wall 10 cm below the bottom of the spring as it hangs unloaded.
-
Solution 2: We can find \(k\) from the test with the bag of water. The force exerted by the bag of water was \((1\text{ kg})(9.8\text{ m/sec}^2) = 9.8\) N\(=k(1 \text{ cm})\text{.}\) So,\begin{equation*} k = \frac{9.8 \frac{\text{kg}\cdot\text{m}}{\text{sec}^2}}{0.01\text{ m}} = 980 \frac{\text{kg}}{\text{sec}^2} \end{equation*}If we hang 10 kg from the spring, gravity exerts a force of \((10 \text{ kg})(9.8 \text{ m/sec}^2) = 98 \frac{\text{kg}\cdot\text{m}}{\text{sec}^2}\text{.}\) This will be matched by the spring with a force of \(kx\) newtons, where \(k\) is the spring constant and \(x\) is the amount stretched.\begin{align*} kx&=98 \frac{\text{kg}\cdot\text{m}}{\text{sec}^2}\\ \left(980 \frac{\text{kg}}{\text{sec}^2}\right)\left(x\text{ m}\right)&=98 \frac{\text{kg}\cdot\text{m}}{\text{sec}^2}\\ x&=\frac{1}{10}\text{ m} = 10\text{ cm} \end{align*}So, we should put the mark at 10 cm below the natural length of the spring.
2.1.2.6.
Solution.
2.1.2.7. (✳).
Solution.
2.1.2.8.
Solution.
- Since \(\frac{c}{\ell-x}\) is measured in newtons, and \(\ell\) and \(x\) (and therefore \(\ell-x\)) are measured in metres, the units of \(c\) are newton-metres, i.e. joules.
- Following Definition 2.1.1, the work done compressing the air is\begin{equation*} W = \int_1^{1.5}F(x)\,\dee{x} \end{equation*}where \(F(x)\) is the amount of force applied when the plunger is \(x\) metres past its natural position. The amount of force applied is equal in magnitude to the amount of force supplied by the tube: \(\frac{c}{\ell-x}\) N. Note \(\ell\) and \(c\) are constants. We can guess the antiderivative, or use the substitution \(u=\ell-x\text{,}\) \(\dee{u}=-\dee{x}\text{.}\)\begin{align*} W &=\int_1^{1.5} \frac{c}{\ell-x}\,\dee{x} = \big[-c\log|\ell-x|\big]_1^{1.5}\\ &=-c\big[\log|\ell-1.5| - \log|\ell-1|\big]\\ &=-c\log\left(\frac{\ell-1.5}{\ell-1}\right)\\ &=c\log\left(\frac{\ell-1}{\ell-1.5}\right)\text{ J} \end{align*}Note that, because \(\ell \gt 1.5\text{,}\) the argument of logarithm is positive, so we don’t need the absolute value signs. Furthermore, \(\ell-1 \gt \ell-1.5\text{,}\) so \(\frac{\ell-1}{\ell-1.5} \gt 1\text{,}\) hence \(\log\left(\frac{\ell-1}{\ell-1.5}\right) \gt 0\text{.}\)
2.1.2.9. (✳).
Solution.
2.1.2.10. (✳).
Solution.
2.1.2.11. (✳).
Solution.
2.1.2.12.
Solution.
- The volume of water in the layer is \(3\dee{y}\) m\(^3\) (since the cross-section has area 3 m\(^3\)).
- One cubic metre is equal to \(100^3\) cubic centimetres. So, the mass of water in one cubic metre is \(\dfrac{100^3}{1000} = 1\,000\) kg.
- Therefore, the mass of water in our layer is (\(3\,000\,\dee{y}\)) kg.
- The force of gravity acting on it is \((-9.8 \times 3\,000\,\dee{y})\) N, so we need to pump with a compensating force of \((9.8 \times 3\,000\,\dee{y})\) N.
- The water needs to be pumped a distance of \(1-y\) metres.
- So, the work required to pump out the thin layer of water at height \(y\) is \((9.8 \times 3\,000\times(1-y)\,\dee{y})\) J.
2.1.2.13. (✳).
Solution.
- has side length \(3-z\) m and hence
- has area \((3-z)^2\) m\(^2\) and hence
- has volume \((3-z)^2\,\dee{z}\) m\(^3\) and hence
- has mass \(8000 (3-z)^2\,\dee{z}\) kg and hence
- is subject to a gravitational force of \(9.8\times 8000 (3-z)^2\,\dee{z}\) N and hence
- requires work \(9.8\times 8000 (2+z)(3-z)^2\,\dee{z}\) J to raise it from \(2\) m below ground level to \(z\) m above ground level.
2.1.2.14.
Solution.
2.1.2.15.
Solution.
- the rope that remains to be lifted has length \(4-y\text{,}\) and so it has mass \(\frac{M}{4}(4-y) \) kg,
- and the firewood still has mass 10 kg.
- The remaining rope and the wood are subject to a downward gravitational force of magnitude \(\underbrace{\left[\frac{M}{4}(4-y)+10\right]}_{\text{mass}}\times\, 9.8\quad\) N.
- So, to raise the firewood from height \(y\) to height \((y+\dee{y})\text{,}\) we need to apply a compensating upward force of \(\left[\frac{M}{4}(4-y)+10\right]\times 9.8\) through distance \(\dee{y}\text{.}\) This takes work \(\left[\frac{M}{4}(4-y)+10\right]\times 9.8\,\dee{y}\) J.
2.1.2.16.
Solution.
-
Solution 1: In this solution, we consider the work on the rope separately from work on the weight, and we imagine lifting a tiny piece of rope the entire distance to the window.The weight has a mass of 5 kg, and is lifted a distance of 5 m to the window. The force of gravity acting on the weight is \((5 \text{ kg})(9.8 \text{ m/sec}^2) = 49\) N, so the work to lift it 5 metres is \((49\text{ N})(5\text{ m}) = \textcolor{red}{245}\text{J}\text{.}\)The density of the rope is \(\frac{1}{10}\) kg/m. A tiny piece of rope of length \(\dee{y}\text{,}\) hanging \(y\) metres from the window, has mass (\(\frac{1}{10}\,\dee{y}\)) kg, and needs to be lifted \(y\) metres. So, the force of gravity acting on the piece of rope is \((\frac{1}{10}\,\dee{y} \text{ kg})\left(9.8\text{ m/sec}^2\right) = 0.98\,\dee{y}\) N, and the work to pull it up to the window is \((0.98y\,\dee{y})\) J. So, the total work to pull up the rope is\begin{equation*} \int_0^{10} 0.98y\,\dee{y} = 0.98\left[\frac{y^2}{2}\right]_0^{10}=\textcolor{blue}{49\text{ J}} \end{equation*}All together, the work to pull up the rope with the weight is \(\textcolor{red}{245}+\textcolor{blue}{49}=294\) J.
-
Solution 2: In this solution, we consider the work on the rope together with the weight, and we imagine lifting the remaining rope a tiny distance to the window.Suppose \(y\) metres of the rope have been pulled in, and \(0 \le y \le 5\) (shown on the left, below). Then the remaining rope has length \(10-y\text{,}\) and contains the weight, so the mass remaining to be pulled up is \(\underbrace{\frac{1}{10}(10-y)}_{\text{rope}} + \underbrace{\vphantom{\frac{1}{10}}5}_{\text{weight}} = 6-\frac{y}{10}\) kg. Then the force of gravity acting on the dangling rope and weight is \((9.8 \text{ m/sec}^2)((6-\frac{y}{10})\text{ kg}) =\left(58.8-0.98y\right)\) N. The work needed to lift this rope \(\dee{y}\) metres is \(\textcolor{red}{\left(58.8-0.98y\right)\dee{y} J}\text{.}\)Now, suppose \(y\) metres of the rope have been pulled in, and \(5 \lt y \le 10\) (shown above, right). Then the remaining rope has length \(10-y\text{,}\) but does not contain the weight, so the mass remaining to be pulled up is \(\frac{1}{10}(10-y) = 1-\frac{y}{10}\) kg. Then the force of gravity acting on the dangling rope is \((9.8 \text{ m/sec}^2)((1-\frac{y}{10})\text{ kg}) =\left(9.8-0.98y\right)\) N. The work needed to lift this rope \(\dee{y}\) metres is \(\textcolor{blue}{\left(9.8-0.98y\right)\dee{y} J}\text{.}\)All together, the work needed to lift the rope is\begin{align*} W \amp= \int_0^{10} F(y) \dee{y} \\ \amp= \textcolor{red}{\int_0^5 \left(58.8-0.98y\right)\dee{y}} + \textcolor{blue}{\int_5^{10}\left(9.8-0.98y\right)\dee{y}}\\ &=\left[58.8y - 0.49y^2\right]_0^{5} + \left[9.8 y - 0.49 y^2\right]_5^{10}\\ &=294 \text{ J} \end{align*}
2.1.2.17.
Solution.
-
The frictional force is \(\mu \times m \times g = 0.4 \left(10 \text{ kg} \right)\left(9.8 \ \frac{\text{m}}{\text{sec}^2}\right) = 39.2\ \frac{\text{kg}\cdot\text{m}}{\text{sec}^2} = 39.2\text { N}\text{.}\) Since this constant force acts over a distance of 3 metres, the work is \(3\times 39.2 =117.6 \text{ J} \text{.}\)In the case of a constant force, we don’t need to use an integral, but we could if we wanted:\begin{equation*} W = \int_0^3 39.2 \dee{x} = \big[39.2x\big]_0^3 = 39.2 \times 3 = 117.6 \text{ J}. \end{equation*}
- Since the box is moving at a speed of 1 m/sec, at time \(t\) we can say the box is at position \(t\text{,}\) \(0 \le t \le 3\text{.}\) At position \(t\text{,}\) the mass of the box is \((10-\sqrt{t})\) kg, so the frictional force is \(0.4 \times m \times g = 0.4\left(10-\sqrt{t} \text{kg}\right)\left(9.8 \frac{\text{m}}{\text{sec}^2}\right) = 3.92(10-\sqrt t) \text{N}\text{.}\) Now that we know the force, to find the work we simply integrate, following Definition 2.1.1:\begin{align*} W &= \int_0^3 3.92(10-\sqrt t)\dee{t} = 3.92\left[10t-\frac{2}{3}t^{3/2}\right]_0^3\\ &=3.92\left[30-\frac{2}{3}\sqrt{3}^3\right]=3.92\left[30-2\sqrt{3}\right]\approx 104\text{ J} \end{align*}
2.1.2.18.
Solution.
2.1.2.19.
Solution.
- Spring constant: The car’s mass of 2000 kg compresses the struts 2 cm past their natural length. The force of the car under gravity is \((2000 \text{ kg})\times(9.8 \text{ m/sec}^2) = 19\,600 \) N. This force is exactly the same as that exerted by the spring, \(k(0.02 \text{ m})\text{.}\) So, \(k = 980\,000\) N/m.
- Work done by spring: The spring can safely compress 20 cm. So, the amount of work done by the spring compressing that far gives us the maximum amount of work the spring can safely do. While the car is falling, the spring is at its natural length, so the work done to compress it to 20 cm (0.2 m) shorter is:\begin{equation*} \int_0^{0.2} kx\,\dee{x} = \int_0^{0.2} 980\,000 x\,\dee{x} =\big[490\,000 x^2 \big]_0^{0.2}=19\,600 \text{ J} \end{equation*}
- Change in kinetic energy: When the car first hits the pavement, it’s falling at \(4\) m/sec, so it has kinetic energy \(\frac{1}{2}(2100\text{ kg})(4\text{ m/sec})^2 = 16\,800\) J. When the car compresses the springs as far as they go and it starts to rebound, it has kinetic energy 0, since its instantaneous velocity is zero. So, the change in kinetic energy is \(16\,800\) J.
2.1.2.20.
Solution.
2.1.2.21. (✳).
Solution.
- has radius \(\sqrt{3^2-x^2}\) m (by Pythagoras) and hence
- has cross-sectional area \(\pi (9-x^2)\) m\(^2\) and hence
- has volume \(\pi (9-x^2)\,\dee{x}\) m\(^3\) and hence
- has mass \(1000 \pi (9-x^2)\,\dee{x}\) kg and hence
- is subject to a gravitational force of \(9.8\times 1000 \pi (9-x^2)\,\dee{x}\) N and hence
- requires work \(9800 \pi (9-x^2)(x+4)\,\dee{x}\) J to raise it to the spout. (It has to be raised \(x\) m to bring it to the height of the centre of the sphere, then \(3\) m more to bring it to the top of the sphere, and finally \(1\) m more to bring it to the spout.)
2.1.2.22.
Solution.
-
Solution 1: Let’s consider the work involved in lifting up a small section of cable, with length \(\dee{y}\text{,}\) distance \(y\) from the bottom end of the cable.The distance this section must travel is \((5-y)\) metres, so if its mass is \(M(y)\text{,}\) then the work involved is\begin{equation*} W = \int_0^5 9.8 \times (5-y)\times M(y) \end{equation*}So, we need to find \(M(y)\text{.}\) The length of the section of cable is \(\dee{y}\text{,}\) and its distance from the end of the cable is \(y\text{,}\) so the mass of the section is \((10-y)\,\dee{y}\text{.}\) Therefore,\begin{align*} W &= \int_0^5 9.8 \times (5-y)\times M(y)\\ &= \int_0^5 9.8 \times (5-y)\times (10-y)\,\dee{y}\\ &=9.8\int_0^5 \big(50-15y+y^2\big)\,\dee{y}\\ &=9.8\left[50y-\frac{15}{2}y^2+\frac{1}{3}y^3\right]_0^5\\ &=\frac{6125}{6} = 1020\tfrac{5}{6}\ \text{J} \end{align*}
-
Solution 2: Alternately, we can continue to use the basic method of Example 2.1.6 in the text, noticing that the density of the cable is no longer constant.Let’s consider pulling the cable up a tiny distance of \(\dee{y}\) metres, after we have already lifted it \(y\) metres (so \((5-y)\) metres of the cable is still in the hole).If \(R(y)\) is the mass of the remaining cable (in kg), then the force of gravity is \(-9.8 \times R(y)\text{,}\) so the work done is \(9.8 \times R(y) \times \dee{y}\text{.}\) Once we find \(R(y)\text{,}\) we can calculate the total work done:\begin{equation*} W = \int _0 ^ 5 9.8 \times R(y) \times \dee{y} \tag{$*$} \end{equation*}As given in the question statement, the density of the cable is \((10-x)\) kg/m, where \(x\) is the distance from the bottom end of the cable. Consider a tiny section of cable \(x\) metres from the bottom end, of length \(\dee{x}\text{.}\)The mass of this tiny section is \(\left(10-x\ \frac{\text{kg}}{\text{m}}\right) \times \left(\dee{x}\ \text{m}\right) = (10-x)\dee{x}\) kg. The section of cable dangling is the last \((5-y)\) metres of cable. So, the combined mass of the section of cable dangling, after we’ve already pulled up \(y\) metres of it, is\begin{equation*} R(y) = \int_0^{5-y} (10-x)\dee{x} =\left[10x - \frac{1}{2}x^2\right]_0^{5-y}=\frac{75}{2} -5y -\frac{1}{2} y^2 \end{equation*}Now we can calculate the total work involved in pulling up the entire cable, using equation (\(*\)).\begin{align*} W &= \int _0 ^ 5 9.8 \times R(y) \times \dee{y}\\ &=\int _0 ^ 5 9.8 \times \left(\frac{75}{2} -5y -\frac{1}{2}y^2\right) \times \dee{y}\\ &=9.8\left[\frac{75}{2}y-\frac{5}{2}y^2-\frac{1}{6}y^3\right]_0^5\\ &=\frac{6125}{6}=1020\frac{5}{6} \text{ J} \end{align*}
2.1.2.23.
Solution.
-
The force of the depends on depth, which varies. So, consider a thin rectangle of the plunger at depth \(y\text{,}\) with height \(\dee{y}\) and width 1 m (the width of the entire plunger). Let the area of this rectangle be \(\dee{A}\text{.}\)The area of this rectangle is \(1\dee{y}\) m\(^2\text{,}\) so the force of the water acting on it is \(F=P\cdot\dee{A} = \underbrace{\left(9800 \tfrac{\text{N}}{\text{m}^3}\right)}_c\underbrace{\left(y\text{ m}\right)}_d \underbrace{\left(\dee{y} \text{ m}^2\right)}_{\dee A} = 9800y\,\dee{y}\) N.The depth at the top of the plunger is \(y=0\text{.}\) To find the depth at the bottom of the plunger, note that the water has a volume of 3 m\(^3\text{,}\) and is in a rectangular container with base 1 m by 3 m. So, its height is 1 m.The force over the entire plunger, from depth \(y=0\) to \(y=1\text{,}\) is\begin{equation*} \int_0^1 9800y\,\dee{y} = \big[4900 y^2\big]_0^1 = 4900 \text{ N} \end{equation*}
-
Let’s follow our work from part (a), but with the width of the length of the base as \(x\) m.Still, a thin rectangle of plunger has width 1 m and height \(\dee{y}\) m, so it has area \(\dee{y}\) m\(^2\text{.}\) At depth \(y\text{,}\) it has a force from the water of \(9800y\,\dee{y}\) N. This hasn’t changed from (a).Now, let’s consider the depth of the water. The volume of water is 3 m\(^3\text{,}\) and it is in a rectangular container with base 1 m by \(x\) m. So, its depth is \(3/x\) m. Therefore, the force on the entire plunger must be calculated from \(y=0\) to \(y=3/x\text{.}\)\begin{equation*} F(x)=\int_0^{3/x} 9800y\,\dee{y} = \big[4900 y^2\big]_0^{3/x} = \frac{9}{x^2}4900 = \frac{44100}{x^2}\text{ N} \end{equation*}Let’s check that this answer makes sense: \(F(3)=4900\) N, which matches our answer from (a).
- If the force of water acting on the plunger, when the length of the base is \(x\) metres, is given by \(F(x)\text{,}\) then we push the plunger with a force of \(-F(x)\text{.}\) Then the work we’re looking for is\begin{align*} W &= \int_3^1 -F(x)\,\dee{x} = \int_1^3 F(x)\,\dee{x}\\ \end{align*}
\(F(x)\) is exactly what we found in (b): \(F(x) = \frac{44100}{x^2}\text{ N}\text{.}\)
\begin{align*} W&= \int_1^3 F(x)\,\dee{x} = \int_1^3 \frac{44100}{x^2}\,\dee{x} = 44100\left[-\frac{1}{x}\right]_1^3 \\ \amp= 44100\cdot \frac{2}{3} = 29\,400\text{ J} \end{align*}
2.1.2.24.
Solution.
2.1.2.25.
Solution.
2.1.2.26.
Solution.
2.1.2.27.
Solution.
Now, we can calculate the volume of a slice at height \(h\) of thickness \(\dee{h}\text{.}\)
\begin{align*} V(h)&=\frac{2}{3}(2-h)^{3/2}\,\dee{h}\\ \end{align*}The density of the liquid at height \(h\) is \(1000\sqrt{2-h}\) kg/m\(^3\text{,}\) so
\begin{align*} M(h)&=1000\sqrt{2-h}\times \frac{2}{3}(2-h)^{3/2}\,\dee{h}\\ &= \frac{2000}{3}(2-h)^2\dee{h}\\ \end{align*}Now we use (\(*\)) to find the work done pumping out the tank.
\begin{align*} W&=\int_0^19.8(1-h)M(h) = \int_0^1 9.8(1-h)\cdot \frac{2000}{3}(2-h)^2\dee{h}\\ &=\frac{19600}{3}\int_0^1 \big(4-8h+5h^2-h^3\big)\,\dee{h}\\ &=\frac{19600}{3}\left[4h-4h^2+\frac{5}{3}h^3-\frac{1}{4}h^4\right]_0^1\\ &=\frac{19600}{3}\left[4-4+\frac{5}{3}-\frac{1}{4}\right]\\ &=\frac{19600}{3}\times\frac{17}{12}\\ &=\frac{83300}{9} = 9255\tfrac{5}{9}\text{ J} \end{align*}2.1.2.28.
Solution.
- Its final position is \(y\) metres above the centre of the hourglass. That is, it was lifted \(2y\) metres against the force of gravity.
- The layer is shaped like a circle with radius \(y^2+0.01\) and height \(\dee{y}\text{,}\) so its volume is \(\pi\left(y^2+0.01\right)^2\,\dee{y}\) cubic metres.
- To find the mass of the layer, we need to know the density of the sand. Let the volume of sand in the hourglass be \(V\text{.}\) We are given its mass \(M\text{.}\) Then \(\pi\left(y^2+0.01\right)^2\,\dee{y}\) cubic metres has a mass of \(\frac{M}{V}\pi\left(y^2+0.01\right)^2\,\dee{y}\) kilograms.
- So, the force of gravity acting on the layer is \(9.8\frac{M}{V}\pi\left(y^2+0.01\right)^2\,\dee{y}\) N, acting vertically downwards.
- To lift the layer to its final position, we apply a compensating force over a distance of \(2y\) metres, for a total work of \(9.8\frac{M}{V}\pi\left(y^2+0.01\right)^22y\,\dee{y}\) J.
- Since the hourglass has height \(0.2\) m, and exactly half of it is filled with sand, the top layer of sand is exactly at the vertical centre of the hourglass, and the bottom layer of sand is 0.1 metres below.
2.1.2.29.
Solution.
- the denominator \(\big(1-x^4\big)^{3/2}\) is always at least as big as \({\big(1-{\big(\frac{1}{2}\big)}^4\big)}^{3/2}\text{,}\)
- the factor \(x^2\) in the numerator is never bigger than \(\big(\frac{1}{2}\big)^2\text{,}\) and
- the factor \(x^4-3\) in the numerator has magnitude at most \(3\text{,}\)
Using Theorem 1.11.13 with \(M = 2\text{,}\) \(a=0\text{,}\) and \(b=1/2\text{,}\) the error in a midpoint approximation with \(n\) intervals is at most
\begin{align*} \frac{M}{24}\cdot\frac{(b-a)^3}{n^2}&=\frac{2}{24}\cdot \frac{1/8}{n^2} = \frac{1}{96n^2} \end{align*}
2.2 Averages
2.2.2 Exercises
2.2.2.1.
Solution.
2.2.2.2.
Solution.
2.2.2.3.
Solution.
2.2.2.4.
Solution.
- The entire interval has length \(b-a\text{,}\) and we’re cutting it into \(n\) pieces, so the length of one piece (and hence the distance between two consecutive samples) is \(\frac{b-a}{n}\text{.}\)
- The first sample, as given in the question statement, is taken at \(x=a\text{.}\) The second sample, then, is at \(x=a+\frac{b-a}{n}\text{,}\) this third is at \(x=1+2\frac{b-a}{n}\text{,}\) and the fourth is at \(a+3\frac{b-a}{n}\text{.}\)
- The \(y\)-value of the fourth sample is simply \(f\left(a+3\frac{b-a}{n}\right)\text{.}\) Note this is the number we use in our average, not the \(x\)-value.
- Our samples are \(f(a)\text{,}\) \(f\left(a+\frac{b-a}{n}\right)\text{,}\) \(f\left(a+2\frac{b-a}{n}\right)\text{,}\) \(f\left(a+3\frac{b-a}{n}\right)\text{,}\) etc. Since there are \(n\) of them, we divide their sum by \(n\text{.}\) So, the average is:\begin{align*} & \frac{f(a)+f\left(a\!+\!\frac{b-a}{n}\right)+ f\left(a\!+\!2\frac{b-a}{n}\right)+\cdots + f\left(a\!+\!(n-1)\frac{b-a}{n}\right)}{n}\\ &=\frac{1}{n}\bigg[f(a)+f\left(a+\frac{b-a}{n}\right)f\left(a+2\frac{b-a}{n}\right)+\cdots \\ \amp\hskip2.2in+ f\left(a+(n-1)\frac{b-a}{n}\right)\bigg]\\ &= \frac{1}{n}\sum_{i=1}^nf\left(a+(i-1)\frac{b-a}{n}\right)\\ \end{align*}As \(n\) gets larger and larger, using the definition of a definite integral, this expression gets closer and closer to \(\frac{1}{b-a}\int_a^bf(x)\,\dee{x}\text{.}\) This is one way of justifying our definition of an average of a function on an interval.
Remark: if we multiply and divide by \(b-a\text{,}\) we see this expression is equivalent to a left Riemann sum, divided by the length of our interval.
\begin{align*} &= \frac{1}{b-a}\sum_{i=1}^nf\left(a+(i-1)\frac{b-a}{n}\right)\frac{b-a}{n}\\ &= \frac{1}{b-a}\sum_{i=1}^nf\left(a+(i-1)\De x\right)\De x \end{align*}
2.2.2.5.
Solution.
- Yes, the average of \(f(x)\) is less than or equal to the average of \(g(x)\) on \([0,10]\text{.}\) The reason is that, if \(f(x) \leq g(x)\) for all \(x\) in \([0,10]\text{,}\) then:\begin{equation*} \frac{1}{10}\int_0^{10}f(x)\,\dee{x} \leq \frac{1}{10}\int_0^{10}g(x)\,\dee{x}. \end{equation*}
-
There is not enough information to tell. It’s certainly possible: for instance, take \(f(x)=0\) and \(g(x)=1\) for all \(x\) in \([0,10]\text{.}\) Then \(f(x) \leq g(x)\) and the average of \(f(x)\) is 0, which is less than 1, the average of \(g(x)\text{.}\)However, consider \(f(x) = \begin{cases} 100 & \text{ if } 0 \le x \le 0.01\\ 0 & \text{ else } \end{cases}\) and \(g(x) = 0\text{.}\) Then \(f(x) \le g(x)\) for all \(x\) in \([0.01,10]\text{,}\) but the average of \(f(x)\) is \(0.1\text{,}\) while the average of \(g(x)\) is 0.
2.2.2.6.
Solution.
2.2.2.7. (✳).
Solution.
2.2.2.8. (✳).
Solution.
To antidifferentiate, we use integration by parts with \(u=\log x\) and \(\dee{v}=x^2\,\dee{x}\text{,}\) hence \(\dee{u}=\frac{1}{x}\,\dee{x}\) and \(v=\frac{1}{3}x^3\text{.}\)
\begin{align*} \frac{1}{e-1}\int_1^e x^2\log x\dee{x}&= \frac{1}{e-1}\left(\left[\frac{1}{3}x^3\log x\right]_1^e - \int_1^e \frac{1}{3}x^2\,\dee{x} \right)\\ &= \frac{1}{e-1}\bigg[\frac{x^3}{3}\log x-\frac{x^3}{9}\bigg]_{x=1}^{x=e}\\ &=\frac{1}{e-1}\bigg[\frac{e^3}{3}-\frac{e^3}{9}+\frac{1}{9}\bigg]\\ &=\frac{1}{e-1}\bigg[\frac{2}{9}e^3+\frac{1}{9}\bigg] \end{align*}2.2.2.9. (✳).
Solution.
2.2.2.10. (✳).
Solution.
2.2.2.11. (✳).
Solution.
Setting \(x=4\text{,}\) we see \(80=8A\text{,}\) so \(\textcolor{red}{A=10}\text{.}\) Setting \(x=-4\text{,}\) we see \(80=8B\text{,}\) so \(\textcolor{red}{B=10}\text{.}\)
\begin{align*} \frac{1}{3}\int_0^3 \frac{80}{16-x^2}\,\dee{x} &=\frac{1}{3}\int_0^3 \frac{80}{(4-x)(4+x)}\,\dee{x}\\ \amp=\frac{1}{3}\int_0^3 \Big[\frac{\textcolor{red}{10}}{4-x}+\frac{\textcolor{red}{10}}{4+x}\Big]\,\dee{x}\\ &=\frac{1}{3}\int_0^3 \Big[-\frac{10}{x-4}+\frac{10}{4+x}\Big]\,\dee{x}\\ &=\frac{10}{3}\Big[-\log|x-4|+\log|x+4|\Big]_0^3\\ &=\frac{10}{3}\log\Big|\frac{x+4}{x-4}\Big|\bigg|_0^3 =\frac{10}{3}[\log 7-\log 1]\\ &=\frac{10}{3}\log 7\quad\text{ degrees Celsius} \end{align*}2.2.2.12. (✳).
Solution.
To integrate, we use the substitution \(u=\log x\text{,}\) \(\dee{u}=\frac{1}{x}\,\dee{x}\text{.}\) Then the limits of integration become 0 and 1, respectively.
\begin{align*} \frac{1}{e-1}\int_1^e\frac{\log x}{x}\dee{x}& =\frac{1}{e-1}\int_0^1u\dee{u} =\frac{1}{e-1}\bigg[\frac{u^2}{2}\bigg]_0^1 =\frac{1}{2(e-1)} \end{align*}2.2.2.13. (✳).
Solution.
2.2.2.14.
Solution.
Since \(\frac{8760}{12} = 730\text{,}\) which is even, \(\sin\left(\frac{8760}{12}\pi\right)=\sin(0)=0\text{.}\) Also, \(\sin\left(\frac{8760}{4380}\pi\right)=\sin(2\pi)=0\text{.}\)
\begin{align*} &=400+\textcolor{red}{\frac{5}{876}(0}) + \textcolor{blue}{\frac{5}{219}(0)}\\ &=400\quad\text{ ppm} \end{align*}Note \(t=0\) to \(t=24\) is one complete period for the integrand in red, so the red integral will evaluate to zero. However, \(t=0\) to \(t=24\) is less than one cycle for the integrand in blue, so we expect this will contribute some nonzero quantity to the average.
\begin{align*} &=400+\textcolor{red}{0} + \textcolor{blue}{\frac{200}{24}\left[\frac{4380}{\pi}\sin\left(\frac{t}{4380}\pi\right)\right]_0^{24}}\\ &=400+\color{blue}{\frac{25}{3}\cdot\frac{4380}{\pi}\sin\left(\frac{24}{4380}\pi\right)}\\ &=400+\color{blue}{\frac{25}{3}\cdot\frac{4380}{\pi}\sin\left(\frac{2}{365}\pi\right)}\\ &\approx 400+\color{blue}{199.99}\\ &=599.99 \quad\text{ ppm} \end{align*}2.2.2.15.
Solution.
- The cross-section of \(S\) at \(x\) is a circle with radius \(x^2\text{,}\) so area \(\pi x^4\text{.}\) The average of these values, \(0 \leq x \leq 2\text{,}\) is\begin{equation*} A = \frac{1}{2-0}\int_0^2 \pi x^4\,\dee{x} = \frac{1}{2}\left[\frac{\pi}{5}x^5\right]_0^2=\frac{16\pi}{5} \end{equation*}
- To find the volume of \(S\text{,}\) imagine cutting it into thin circular disks of radius \(x^2\) and thickness \(\dee{x}\text{.}\) The volume of one such disk is \(\pi x^4\,\dee{x}\text{,}\) so the volume of \(S\) is\begin{equation*} \int_0^2 \pi x^4\,\dee{x} = \left[\frac{\pi}{5}x^5\right]_0^2 = \frac{32\pi}{5} \end{equation*}
-
The volume of a cylinder is the product of its base area with its length. A cylinder with circular cross-sections of area \(\frac{16\pi}{5}\) and length 2 has volume \(\frac{32\pi}{5}\text{.}\)Remark: this is the same as the volume of \(S\text{,}\) so the average cross-sectional area of \(S\) tells us the cross-sectional area of a cylinder with the same length and volume as \(S\text{.}\) Compare this to Question 1, where we saw the average value of a function gave the height of a rectangle with the same area as the function over the given interval.
2.2.2.16.
Solution.
- We can see without calculation that the average will be zero, since \(f(x)=x\) is an odd function and \([-3,3]\) is a symmetric interval. Alternately, we can use the definition of an average to calculate\begin{equation*} \frac{1}{6}\int_{-3}^3 x\,\dee{x} = \left[\frac{1}{12}x^2\right]_{-3}^3 = \frac{1}{12}(9-9)=0 \end{equation*}
- Using the definition provided for root mean square:\begin{align*} \text{RMS}&=\sqrt{\frac{1}{6}\int_{-3}^3 x^2\,\dee{x}} = \sqrt{\left[\frac{1}{18}x^3\right]_{-3}^3} = \sqrt{\frac{27}{18}-\frac{-27}{18}} = \sqrt{3} \end{align*}
2.2.2.17.
Solution.
2.2.2.18.
Solution.
- Using Hooke’s law, when the spring is stretched (or compressed) \(f(t)\) metres past its natural length, the force exerted is \(kf(t)\text{,}\) where \(k\) is the spring constant. In this case, the force is\begin{equation*} F(x) = (3\text{ N/cm})(f(t)\text{ cm}) = 3 \sin\left(t\pi\right)\text{ N} \end{equation*}
-
Our interval encompasses three full periods of sine, so the average will be zero.Alternately, we can compute, using the definition of an average:\begin{align*} \text{Avg}&=\frac{1}{6}\int_0^6 3\sin(t\pi)\,\dee{t}= \frac{1}{6}\left[-\frac{3}{\pi}\cos(t\pi)\right]_0^6\\ \amp=\frac{1}{2\pi}\left[\cos 0 - \cos (6\pi)\right]=0 \end{align*}This it doesn’t tell us very much about the “normal” amount of force from the spring during our time period. It only tells us that force in one direction at is “cancelled out” by force in the opposite direction at another time.
- Using the definition given for root mean square,\begin{align*} \text{RMS}&=\sqrt{\frac{1}{6}\int_0^6 \left(3\sin(t\pi)\right)^2\,\dee{t}} = \sqrt{\frac{3}{2}\int_0^6 \sin^2(t\pi)\,\dee{t}}\\ & = \sqrt{\frac{3}{4}\int_0^6 \left(1-\cos(2t\pi)\right)\,\dee{t}}\\ &=\sqrt{\frac{3}{4}\left[t - \frac{1}{2\pi}\sin(2t\pi)\right]_0^6}\\ &=\sqrt{\frac{3}{4}\left[6 - \frac{1}{2\pi}\sin(12\pi)-0\right]}\\ &=\sqrt{\frac{3}{4}(6)}=\frac{3}{\sqrt2}\approx 2.12 \end{align*}
2.2.2.19. (✳).
Solution.
2.2.2.20.
Solution.
- Using the definition of an average,\begin{equation*} A=\frac{1}{1-0}\int_0^1 e^t\,\dee{t} = e-1 \end{equation*}
-
Since \(s(t)-A = e^t-e+1\text{,}\) its average on \([0,1]\) is\begin{align*} \frac{1}{1-0}\int_0^1 \left(e^t-e+1\right)\,\dee{t} \amp= \left[e^t -et+t\right]_0^1 \\ \amp= (e-e+1)-(1)=0 \end{align*}Remark: what’s happening here is that the average difference between \(s(t)\) and \(A\) is zero, because the values of \(s(t)\) that are larger than \(A\) (and give a positive value of \(s(t)-A\)) exactly cancel out the values of \(s(t)\) that are smaller than \(A\) (and give a negative value of \(s(t)-A\)). However, knowing how far the average value is from our calculated average is a reasonable thing to measure. That’s where (c) comes in.
-
Using the definition of an average, the quantity we want is:\begin{equation*} \frac{1}{1-0}\int_0^1\left| e^t-e+1\right|\,\dee{t} \end{equation*}To deal with the absolute value, we consider the integral over two intervals: one where \(e^t-e+1\) is positive, and one where it’s negative. To decide where to break the limits of integration, notice \(e^t-e+1 \gt 0\) exactly when \(e^t \gt e-1\text{,}\) so \(t \gt \log(e-1)\text{.}\)\begin{align*} &\frac{1}{1-0}\int_0^1\left| e^t-e+1\right|\,\dee{t}\\ \amp=\int_0^{\log(e-1)}|\underbrace{e^t-e+1}_{\text{negative}}|\,\dee{t}+\int_{\log(e-1)}^1|\underbrace{e^t-e+1}_{\text{positive}}|\,\dee{t}\\ &=\int_0^{\log(e-1)}\big(-e^t+e-1\big)\,\dee{t}+\int_{\log(e-1)}^1\big(e^t-e+1\big)\,\dee{t}\\ &=\Big[-e^t+(e-1)t\Big]_0^{\log(e-1)}+ \Big[e^t-(e-1)t\Big]_{\log(e-1)}^1\\ &=\left[-(e-1)+(e-1)\log(e-1)+1\right]\\ \amp\hskip1in+ \left[e-(e-1)-(e-1)+(e-1)\log(e-1)\right]\\ &=4-2e+2(e-1)\log(e-1)\\ &\approx 0.42 \end{align*}Remark: what we just measured is how far \(s(t)\) is, on average, from \(A\text{.}\) We had to neglect whether \(s(t)\) was above or below \(A\text{,}\) because (as we saw in (b)) the values above \(A\) “cancel out” the values below \(A\text{.}\) That’s where the absolute value came in.Knowing how well most of your function’s values match the average is an important measure, but dealing with absolute values can be a little clumsy. Therefore, the variance of a function squares the differences, rather than taking their absolute value. (In our example, that means looking at \((s(t)-A)^2\text{,}\) rather than \(|s(t)-A|\text{.}\)) To compensate for the change in magnitude involve in squaring, the standard deviation is the square root of the variance. These are two very commonly used measures of how similar a function is to its average. Compare standard deviation to root-square-mean voltage from Example 2.2.6 and Questions 16 to 18.
2.2.2.21.
Solution.
-
Neither: the average of both these functions is zero. We saw this with a particular function in Question 20 (b), but it’s actually true in general. It’s a quick calculation to prove.The average of \(f(x)-A\) is:\begin{align*} \frac{1}{4-0}\int_0^4 \big(f(x)-A\big)\,\dee{x}&=\underbrace{\frac{1}{4}\int_0^4f(x)\,\dee{x}}_{A} - A=A-A=0 \end{align*}Similarly, the average of \(g(x)-A\) is:\begin{align*} \frac{1}{4-0}\int_0^4 \big(g(x)-A\big)\,\dee{x}&=\underbrace{\frac{1}{4}\int_0^4g(x)\,\dee{x}}_{A} - A=A-A=0 \end{align*}
-
The function \(|f(x)-A|\) tells us how far \(f(x)\) is from \(A\text{,}\) without worrying whether \(f(x)\) is larger or smaller. Looking at our graph, for most values of \(x\) in \([0,4]\text{,}\) \(f(x)\) is quite far away from \(A\text{,}\) so \(|f(x)-A|\) is usually a large, positive quantity.By contrast, \(|g(x)-A|\) is a small positive quantity for most values of \(x\text{.}\) The function \(g(x)\) is quite close to \(A\) for all values of \(x\) in \([0,4]\text{.}\)So, since \(|g(x)-A|\) generally has much smaller values than \(|f(x)-A|\text{,}\) the average of \(|f(x)-A|\) on \([0,4]\) will be larger than the average of \(|g(x)-A|\) on \([0,4]\text{.}\)As discussed in Question 20(c), the average of \(|f(x)-A|\) is a measure of how closely \(f(x)\) resembles its average. We see from the graph that \(f(x)\) doesn’t resemble the constant function \(y=A\) much at all, while \(g(x)\) seems much more similar to the constant function \(y=A\text{.}\)This kind of measure — how similar a function is to its average — is also the idea behind the root square mean.
2.2.2.22.
Solution.
2.2.2.23.
Solution.
2.2.2.24.
Solution.
2.2.2.25.
Solution.
2.2.2.26.
Solution.
- The function \(A(x)\) only gives us information about an integral when one limit of integration is zero. We can get around this by using properties of definite integrals from Section 1.2 to break our integral into two integrals, each of which has 0 as one limit of integration. So, we find the average of \(f(t)\) on \([a,b]\) as follows:\begin{align*} \amp\frac{1}{b-a}\int_a^b f(t)\,\dee{t}= \frac{1}{b-a}\left(\int_a^0 f(t)\,\dee{t} + \int_0^b f(t)\,\dee{t}\right)\\ &\hskip0.5in=\frac{1}{b-a}\left(-\int_0^a f(t)\,\dee{t} + \int_0^b f(t)\,\dee{t} \right)\\ &\hskip0.5in=\frac{1}{b-a}\left(-a\cdot\underbrace{\frac{1}{a}\int_0^a f(t)\,\dee{t}}_{A(a)} + b\cdot\underbrace{\frac{1}{b}\int_0^b f(t)\,\dee{t}}_{A(b)} \right)\\ &\hskip0.5in=\frac{1}{b-a}\left(-aA(a)+bA(b) \right) = \frac{bA(b) - aA(a)}{b-a} \end{align*}
-
From the definition of \(A(x)\text{,}\) we know\begin{equation*} A(x) = \frac{1}{x}\int_0^x f(t)\,\dee{t} \end{equation*}That is,\begin{equation*} xA(x) =\int_0^x f(t)\,\dee{t} \end{equation*}To find \(f(x)\text{,}\) we differentiate both sides. For the left side, we use the product rule; for the right side, we use the Fundamental Theorem of Calculus part 1.\begin{equation*} A(x)+xA'(x) = f(x) \end{equation*}So, \(f(t)=A(t)+tA'(t)\text{.}\)
2.2.2.27.
Solution.
- One of many possible answers: \(f(x) = \begin{cases} -1&\text{ if } x \leq 0\\ 1&\text{ if } x \gt 0 \end{cases}\text{.}\)
- No such function exists.
- Note 1: Suppose \(f(x) \gt 0\) for all \(x\) in \([-1,1]\text{.}\) Then \(\frac{1}{2}\int_{-1}^1f(x)\,\dee{x} \gt \frac{1}{2}\int_{-1}^10\,\dee{x} =0 \text{.}\) That is, the average value of \(f(x)\) on the interval \([-1,1]\) is not zero — it’s something greater than zero.
- Note 2: Suppose \(f(x) \lt 0\) for all \(x\) in \([-1,1]\text{.}\) Then \(\frac{1}{2}\int_{-1}^1f(x)\,\dee{x} \lt \frac{1}{2}\int_{-1}^10\,\dee{x} =0 \text{.}\) That is, the average value of \(f(x)\) on the interval \([-1,1]\) is not zero — it’s something less than zero.
So, if the average value of \(f(x)\) is zero, then \(f(x)\ge 0\) for some \(x\) in \([-1,1]\text{,}\) and \(f(y) \le 0\) for some \(y \in [-1,1]\text{.}\) Since \(f\) is a continuous function, and 0 is between \(f(x)\) and \(f(y)\text{,}\) by the intermediate value theorem (see the CLP-1 text) there is some value \(c\) between \(x\) and \(y\) such that \(f(c)=0\text{.}\) Since \(x\) and \(y\) are both in \([-1,1]\text{,}\) then \(c\) is as well. Therefore, no function exists as described in the question.
2.2.2.28.
Solution.
If \(\int_0^\infty f(t) \,\dee{t}\) converges, then this limit is 0, and the statement is true. So, suppose it does not converge. Since \(f(x)\) is positive, that means \(\lim\limits_{x \to \infty} \int_0^x f(t)\,\dee(t) = \infty\text{,}\) so we can use l’Hôpital’s rule. To differentiate the numerator, we use the Fundamental Theorem of Calculus part 1.
\begin{align*} \lim_{x \to \infty}A(x)&=\lim_{x \to \infty}\frac{\int_0^x f(t)\,\dee{t}}{x} = \lim_{x \to \infty}\frac{f(t)}{1}=0 \end{align*}2.2.2.29.
Solution.
2.3 Centre of Mass and Torque
2.3.3 Exercises
2.3.3.1.
Solution.
2.3.3.2.
Solution.
2.3.3.3.
Solution.
2.3.3.4.
Solution.
2.3.3.5.
Solution.
2.3.3.6.
Solution.
- The volume of water in Tank A is \(\frac{4}{3}\pi(1)^3 = \frac{4}{3}\pi\) cubic metres.
- The mass of water is \(\frac{4000}{3}\pi\) kg.
- By symmetry, the centre of mass of the water when it fills Tank \(A\) is exactly in the centre of the sphere, at height \(\bar y_1=4\) metres above the ground (one metre above the bottom of Tank \(A\text{,}\) which is three metres above the ground).
- When the water is entirely in Tank \(B\text{,}\) its height is \(\frac{2}{3}\pi\) metres. (The base of Tank \(B\) has area 2 m\(^2\text{,}\) and the volume of water is \(\frac{4}{3}\pi\) m\(^3\text{.}\)) By symmetry, the centre of mass is exactly halfway up, at height \(\bar y_2=\frac{1}{3}\pi\) metres.
- So, the point mass in our model is moved from \(\bar y_1=4\) to \(\bar y_2=\frac{1}{3}\pi\text{,}\) a distance of \(4-\frac{1}{3}\pi\) metres, by gravity.
- The work involved is:\begin{align*} \left(\frac{4000}{3}\pi \text{kg}\right)\times\left(4-\frac{1}{3}\pi \text{m}\right)\times\left(9.8 \frac{\text{m}}{\text{sec}^2}\right) \amp= \frac{39200\pi}{9}(12-\pi)\\ \amp\approx 121,212 \text{J} \end{align*}
2.3.3.7.
Solution.
- A thin slice of \(S\) at position \(x\) has height \(\frac{1}{x}\text{,}\) so if its width is \(\dee{x}\text{,}\) its area is \(\frac{1}{x}\,\dee{x}\text{.}\)
- A small piece of \(R\) at position \(x\) has density \(\frac{1}{x}\text{,}\) so if its length is \(\dee{x}\text{,}\) its mass is \(\frac{1}{x}\,\dee{x}\text{.}\)
- Adding up all our tiny slices from (a) gives us the total area of \(S\text{:}\)\begin{equation*} \int_1^3 \frac{1}{x}\,\dee{x} = \log 3 \end{equation*}
- Adding up all our tiny pieces from (b) gives us the total mass of \(R\text{:}\)\begin{equation*} \int_1^3 \frac{1}{x}\,\dee{x}=\log 3 \end{equation*}
- Using Equation 2.3.5, the \(x\)-coordinate of the centroid of \(S\) is\begin{equation*} \frac{\int_1^3 x\cdot\frac{1}{x}\,\dee{x}}{\int_1^3\frac{1}{x}\,\dee{x}} = \frac{\int_1^3 1\,\dee{x}}{\log 3} = \frac{2}{\log 3} \end{equation*}
- Using Equation 2.3.4, the centre of mass of \(R\) is\begin{equation*} \frac{\int_1^3 x\cdot\frac{1}{x}\,\dee{x}}{\int_1^3\frac{1}{x}\,\dee{x}} = \frac{\int_1^3 1\,\dee{x}}{\log 3} = \frac{2}{\log 3} \end{equation*}
2.3.3.8.
Solution.
-
If we chop \(R\) into \(n\) pieces, each piece has length \(\frac{b-a}{\vphantom{\frac12}n}\text{.}\) Then our \(i\)th cut is at position \(a+i\left(\frac{b-a}{n}\right)\text{,}\) so our \(i\)th piece runs from \(a+(i-1)\left(\frac{b-a}{n}\right)\) to \(a+i\left(\frac{b-a}{n}\right)\text{.}\) The approximation of the mass of this piece comes from the density at its midpoint,\begin{align*} m_i\amp=\frac{\left[a+(i-1)\left(\frac{b-a}{n}\right)\right]+\left[a+i\left(\frac{b-a}{n}\right)\right]}{2} \\ \amp= a+(i-\tfrac12)\left(\frac{b-a}{n}\right) \end{align*}So, the \(i\)th piece has length \(\frac{b-a}{\vphantom{\frac12}n}\text{,}\) with approximate density \(\rho\left(m_i\right)=\rho\left(a+(i-\tfrac12)\left(\frac{b-a}{n}\right)\right)\text{.}\) We approximate that the \(i\)th piece has mass \(\left(\frac{b-a}{n}\right)\cdot\rho\left(m_i\right)\) and position \(m_i\text{.}\) Using Equation 2.3.1, the centre of mass of \(R\) is approximately at position:\begin{align*} \bar x_n &= \frac{\sum\limits_{i=1}^n\text{(mass of $i$th piece)}\times\text{(position of $i$th piece)}}{\sum\limits_{i=1}^n\text{(mass of $i$th piece)}}\\ &=\frac{\sum\limits_{i=1}^n\left[\frac{b-a}{n}\rho(\textcolor{red}{m_i})\times\textcolor{red}{m_i}\right]}{\sum\limits_{i=1}^n\frac{b-a}{n}\rho(\textcolor{red}{m_i})}\\ &=\frac{\sum\limits_{i=1}^n\left[\frac{b-a}{n}\rho\left(\textcolor{red}{a+\left(i-\tfrac12\right)(\tfrac{b-a}{n})}\right)\times\left( \textcolor{red}{ a+(i-\tfrac12)\left(\tfrac{b-a}{n}\right) } \right)\right]}{\sum\limits_{i=1}^n\frac{b-a}{n}\rho\left(\textcolor{red}{ a+(i-\tfrac12)\left(\tfrac{b-a}{n}\right) } \right)} \end{align*}
-
Remember the definition of a midpoint Riemann sum:\begin{equation*} \int_a^b f(x)\,\dee{x} \approx \sum_{i=1}^n \frac{b-a}{n}\cdot f\left(a+(i-\tfrac12)\left(\frac{b-a}{n}\right)\right) \end{equation*}The numerator of our approximation in part (a) is, therefore, a midpoint Riemann sum of \(\int_a^b \rho(x)\times x\,\dee{x}\text{,}\) and the denominator is a midpoint Riemann sum of \(\int_a^b \rho(x)\,\dee{x}\text{.}\)Using the definition of a definite integral (Definition 1.1.9), we see the limit of the approximation in (a) as \(x\) goes to infinity is\begin{equation*} \bar x=\frac{\int_a^b x\rho(x)\,\dee{x}}{\int_a^b\rho(x)\,\dee{x}} \end{equation*}This gives us the exact centre of mass of our rod.Remark: this is Equation 2.3.4 in the text.
2.3.3.9.
Solution.
-
On the left-most corner of \(S\text{,}\) \(T(x)=B(x)\text{,}\) so the height of \(S\) is zero; that is, the area of a very small vertical strip is very close to zero, so the density of \(R\) is close to 0. As we move closer to the position labeled \(a'\text{,}\) the height of the strips increases, so the areas of the strips increases, so the density of \(R\) increases. Then, between the points labeled \(a'\) and \(b'\text{,}\) the height of \(S\) remains constant, since \(T(x)\) and \(B(x)\) are parallel here, so the areas of the strips of \(S\) remain constant, and the density of \(R\) remains constant. Then, between \(b'\) and \(b\text{,}\) the height of \(S\) decreases, so the area of the strips decrease, so the density of \(R\) decreases.
- At position \(x\text{,}\) the height of \(S\) is \(T(x)-B(x)\text{,}\) so a rectangle with width \(\dee{x}\) and this height would have area \((T(x)-B(x))\,\dee{x}\text{.}\)
- According to our model, the tiny section of \(R\) at position \(x\) with width \(\dee{x}\) has mass \((T(x)-B(x))\,\dee{x}\) (that is, the area of \(S\) over this same tiny interval), so its density is \(\rho(x) = \frac{\text{mass}}{\text{length}} = \frac{(T(x)-B(x))\,\dee{x}}{\dee{x}} = T(x)-B(x)\text{.}\)
-
Imagine \(S\) were a solid, of constant density. The mass of a portion of \(S\) is proportional to the area of that portion. To find the \(x\)-coordinate where the solid would balance, we imagine compressing together the vertical dimension of \(S\) until it’s a rod. That is, we would take a very thin vertical strip of \(S\text{,}\) and turn it into a small segment of a rod, with the same mass. Then the centre of mass of that rod would be exactly the \(x\)-coordinate of the centre of mass of the solid — that is, the \(x\)-coordinate of the centroid of \(S\text{.}\)The compressed rod we form in this way is exactly \(R\) (perhaps multiplied by a constant, to account for the density of \(S\text{,}\) but this doesn’t affect where \(R\) balances). So, the \(x\)-coordinate of the centroid has the same position as the centre of mass of \(R\text{.}\)\begin{align*} & \frac{\int_a^b x\textcolor{red}{\rho(x)}\,\dee{x}}{\int_a^b\textcolor{red}{\rho(x)}\,\dee{x}}\\ \end{align*}
In (c), we found \(\rho(x)=T(x)-B(x)\text{.}\) So, for the solid \(S\) bounded by \(T(x)\) and \(B(x)\) on the interval \([a,b]\text{,}\)
\begin{align*} \bar x&=\frac{\int_a^b x(\textcolor{red}{T(x)-B(x)})\,\dee{x}}{\int_a^b(\textcolor{red}{T(x)-B(x)})\,\dee{x}} \end{align*}Remark: the denominator is the area of \(S\text{.}\) This formula is the same as the formula found in Equation 2.3.5.
2.3.3.10.
Solution.
-
To begin with, we’ll sketch some strips, and put a dot at the centre of mass of each one (its vertical centre).In our model, each of these strips corresponds to a weight on \(R\text{,}\) positioned at its centre of mass (the height of the dot), and with a mass equal to the strip’s area. For the portion of \(S\) with \(a'\le x \le b'\text{,}\) each centre of mass is at a slightly different height, but the areas of the slices are the same. So, the corresponding weights along \(R\) are at different heights, but all have the same mass, as shown below. (Note the rod \(R\) below only contains the weights from the middle of \(S\) — we’ll add the rest later.)For clarity, the diagrams below are zoomed in.By contrast to the slices in the interval \([a',b']\text{,}\) the slices of \(S\) along \([a,a']\) all have the same centre of mass, but different areas. So, there is one position along \(R\) that has a number of weights all stacked on top of one another, of varying masses.The same situation applies to the slices of \(S\) along \([b,b']\text{.}\) So, all together, our rod looks something like this:Remark: if we had sketched the density of \(R\text{,}\) it would have looked something like this:because from our sketch, we see that the density of \(R\text{:}\)
- is 0 at either end,
- is suddenly very high where the blue weights are, and
- is constant and lower between the blue weights.
-
At position \(x\text{,}\) the height of \(S\) is \(T(x)-B(x)\text{,}\) and the width of the strip is \(\dee{x}\text{,}\) so the area of the strip is \((T(x)-B(x))\,\dee{x}\text{.}\)Since the density of \(S\) is uniform, the centre of mass of the strip is halfway up: at \(\dfrac{T(x)+B(x)}{2}\text{.}\)
-
If we cut \(S\) into \(n\) strips, then the strip at position \(x_i\) has area \((T(x_i)-B(x_i))\De x\text{,}\) where \(\De x = \frac{b-a}{n}\text{,}\) and its centre of mass is at height \(\dfrac{T(x_i)+B(x_i)}{2}\text{.}\) So, our approximation of the centre of mass of the rod is:\begin{align*} \bar y_n &=\frac{\sum\limits_{i=1}^n (M_i\times y_i) }{\sum\limits_{i=1}^n M_i}\\ &=\frac{\sum\limits_{i=1}^n \left((T(x_i)-B(x_i))\De x\right)\times\left(\dfrac{T(x_i)+B(x_i)}{2}\right) }{\sum\limits_{i=1}^n (T(x_i)-B(x_i))\De x}\\ &=\frac{\sum\limits_{i=1}^n (T(x_i)^2-B(x_i)^2)\De x }{2\sum\limits_{i=1}^n (T(x_i)-B(x_i))\De x}\\ \end{align*}
We use the definition of a definite integral (Definition 1.1.9) to re-write the limit of the above function.
\begin{align*} \bar y &=\lim_{n \to \infty}\frac{\sum\limits_{i=1}^n (T(x_i)^2-B(x_i)^2)\De x }{2\sum\limits_{i=1}^n (T(x_i)-B(x_i))\De x}\\ &=\frac{\int_a^b \big(T(x)^2-B(x)^2\big)\,\dee{x}}{2\int_a^b\big( T(x)-B(x)\big)\,\dee{x}} \end{align*}Remark: the denominator is twice the area of \(S\text{.}\) This equation for the \(y\)-coordinate of the centroid is the same as the one given in Equation 2.3.5.
2.3.3.11. (✳).
Solution.
2.3.3.12.
Solution.
2.3.3.13.
Solution.
For the numerator, we use the substitution \(u=1+x^2\text{,}\) \(\dee{u}=2x\,\dee{x}\text{.}\)
\begin{align*} &= \frac{\frac{1}{2}\int_{10}^{101} \frac{1}{u}\,\dee{u}}{\Big[\arctan x \Big]_{-3}^{10} } =\frac{\frac{1}{2}\Big[\log u\Big]_{10}^{101}}{\arctan 10 - \arctan(-3)}\\ &=\frac{\Big[\log 101-\log 10\Big]}{2(\arctan 10 + \arctan(3))}=\frac{\log 10.1}{2(\arctan 10 + \arctan(3))}\approx 0.43 \end{align*}2.3.3.14. (✳).
Solution.
2.3.3.15. (✳).
Solution.
- For all \(x\) in its domain, \(T(x) \geq 0\text{.}\) In particular, it’s always the top of our region (so \(T(x)\) is a reasonable name for it), while the bottom is \(B(x)=0\text{.}\)
- \(T(0)=\frac{1}{4}\text{,}\) and \(T(2)=\frac{1}{2\sqrt{3}}\)
-
\(T'(x) = \frac{x}{(16-x^2)^{3/2}}\text{,}\) which is positive on \([0,2]\text{,}\) so \(T(x)\) is increasing.Remark: to see that \(T(x)\) is increasing, we can also just break it into pieces:
- When \(x \ge 0\text{,}\) \(x^2\) is increasing, so
- \(16-x^2\) is decreasing, so
- \(\sqrt{16-x^2}\) is decreasing, so
- \(\frac{1}{\sqrt{16-x^2}}=T(x)\) is increasing.
- \(T''(x)=\frac{2x^2+16}{(16-x^2)^{5/2}}\text{,}\) which is positive, so \(T(x)\) is concave up.
- the right endpoint of a horizontal strip is always \(x=2\text{,}\)
- the left endpoint is determined by \(T(x)\) from \(y=\frac{1}{4}\) to \(y=\frac{1}{2\sqrt{3}}\text{,}\) and
- the left endpoint is \(x=0\) for \(0 \le y \le \frac{1}{4}\text{.}\)
Using the method of partial fractions, we see \(\displaystyle\frac{1}{16-x^2} = \frac{1/8}{4+x}+\frac{1/8}{4-x}\text{.}\)
\begin{align*} &= \frac{1}{2A}\int_0^2 \Big[\frac{1/8}{4+x}+\frac{1/8}{4-x}\Big]\,\dee{x} = \frac{1}{16A}\int_0^2 \Big[\frac{1}{x+4}-\frac{1}{x-4}\Big]\,\dee{x}\\ &= \frac{1}{16A} \Big[\log|x+4|-\log|x-4|\Big]_0^2 = \frac{6}{16\pi} \big[\log6-\log2-\log4+\log 4\big]\\ &=\frac{3\log 3}{8\pi} \end{align*}2.3.3.16. (✳).
Solution.
We use integration by parts with \(u=x\text{,}\) \(\dee{v}=(\cos x - \sin x)\dee{x}\text{;}\) \(\dee{u}=\dee{x}\text{,}\) \(v=\sin x + \cos x\text{.}\)
\begin{align*} &=\frac{1}{A}\left(\Big[x(\sin x + \cos x )\Big]_0^{\pi/4} - \int_0^{\pi/4} (\sin x + \cos x)\,\dee{x}\right)\\ &=\frac{1}{A}\Big[x\sin(x)+x\cos(x)+\cos x -\sin x\Big]_0^{\pi/4}\\ &=\frac{1}{A}\left[\left(\frac{\pi}{4}\cdot\frac{1}{\sqrt2}+\frac{\pi}{4}\cdot\frac{1}{\sqrt{2}}+ \frac{1}{\sqrt2}-\frac{1}{\sqrt2}\right)- 1\right]\\ &=\frac{\frac{\pi}{4}\sqrt{2}-1}{A}=\frac{\frac{\pi}{4}\sqrt{2}-1}{\sqrt{2}-1}\\ \end{align*}Again using Equation 2.3.5,
\begin{align*} \bar y&= \frac{1}{2A}\int_0^{\pi/4} \big(T(x)^2-B(x)^2\big)\,\dee{x} =\frac{1}{2A}\int_0^{\pi/4} \big(\cos^2(x)-\sin^2(x)\big)\,\dee{x}\\ &=\frac{1}{2A}\int_0^{\pi/4} \cos(2x)\,\dee{x} =\frac{1}{2A}\Big[\frac{1}{2}\sin(2x)\Big]_0^{\pi/4} =\frac{1}{4(\sqrt{2}-1)} \end{align*}2.3.3.17. (✳).
Solution.
Although we have a quadratic function underneath a square root, we find an easier method than a trig substitution: the substitution \(u=1+x^2,\dee{u}=2x\,\dee{x}\text{.}\) This changes the limits of integration to \(1+0^2=1\) and \(1+1^2=2\text{,}\) respectively.
\begin{align*} &= \frac{1}{A}\int_1^2\frac{k}{\sqrt{u}}\,\frac{\dee{u}}{2} =\frac{k}{2A}\left[\frac{\sqrt{u}}{1/2}\right]_1^2 =\frac{k}{A}\big[\sqrt{2}-1\big] \end{align*}Since \(k\) and A are a positive constants (hence neither is equal to 0), we can divide both sides by \(k\) and multiply both sides by \(A\text{:}\)
\begin{align*} \sqrt{2}-1&=\frac{k\pi}{8}\\ k&=\frac{8}{\pi}\big[\sqrt{2}-1\big] \end{align*}2.3.3.18. (✳).
Solution.
2.3.3.19. (✳).
Solution.
We can guess the antiderivative in the numerator, or use the substitution \(u=1+x^2\text{,}\) \(\dee{u}=2x\,\dee{x}\text{.}\)
\begin{align*} &=\frac{\half\log(1+x^2)\big|_0^1}{\arctan x\big|_0^1} =\frac{\half\log 2}{\pi/4} =\frac{2}{\pi}\log 2 \approx 0.44127 \end{align*}2.3.3.20. (✳).
Solution.
2.3.3.21. (✳).
Solution.
2.3.3.22.
Solution.
1
Since \(y\) is an odd function, and the domain of integration is symmetric, the first integral evaluates to 0. Since \(y\sin y\) is an even function (recall the product of two odd functions is an even function), we can simplify our limits of integration.
\begin{align*} &=-\frac{2}{\pi}\int_0^{\pi/2} y\sin y\,\dee{y}\\ \end{align*}We use integration by parts with \(u=y\text{,}\) \(\dee{v}=\sin y\,\dee{y}\text{;}\) \(\dee{u}=\dee{y}\text{,}\) \(v=-\cos y\text{.}\)
\begin{align*} &=-\frac{2}{\pi}\left( \big[-y\cos y\big]_0^{\pi/2}+ \int_0^{\pi/2} \cos y\,\dee{y} \right)\\ &=-\frac{2}{\pi} \big[-y\cos y+\sin y\big]_0^{\pi/2}\\ &=-\frac{2}{\pi}\left[(0+1)- 0\right] = -\frac{2}{\pi} \end{align*}2.3.3.23.
Solution.
Using Equation 2.3.5:
\begin{align*} \bar x &=\frac{\int_0^2 x(T(x)-B(x))\,\dee{x}}{A}\\ &= \frac{1}{e^2-5/2}\left[\int_0^1x\left(e^x-0\right)\,\dee{x} + \int_1^2 x\left(e^x-3(x-1)\right)\,\dee{x}\right]\\ &= \frac{1}{e^2-5/2}\left[\int_0^2 xe^x\,\dee{x} - \int_1^2 3x(x-1)\,\dee{x}\right]\\ \end{align*}For the left integral, we use integration by parts with \(u=x\text{,}\) \(\dee{v}=e^x\,\dee{x}\text{;}\) \(\dee{u}=\dee{x}\text{,}\) \(v=e^x\text{.}\)
\begin{align*} &= \frac{1}{e^2-5/2}\left[\left[xe^x\right]_0^2-\int_0^2 e^x\,\dee{x} - 3\int_1^2 (x^2-x)\,\dee{x}\right]\\ &= \frac{1}{e^2-5/2}\left(\left[xe^x-e^x\right]_0^2 - 3\left[\frac{1}{3}x^3-\frac{1}{2}x^2\right]_1^2\right)\\ &= \frac{1}{e^2-5/2}\left((2e^2-e^2)-(-1) - 3\left(\frac{8}{3}-2-\frac{1}{3}+\frac{1}{2}\right)\right)\\ &= \frac{e^2-3/2}{e^2-5/2}\approx 1.2\\ \end{align*}Using Equation 2.3.5 again:
\begin{align*} \bar y &=\frac{\int_0^2\left(T(x)^2-B(x)^2\right)\,\dee{x}}{2A}\\ &=\frac{1}{2(e^2-5/2)}\left[\int_0^1\left(e^{2x} - 0\right)\,\dee{x} + \int_1^2\left(e^{2x} - 9(x-1)^2\right)\,\dee{x}\right]\\ &=\frac{1}{2(e^2-5/2)}\left[\int_0^2 e^{2x}\,\dee{x} - \int_1^2 9(x-1)^2\,\dee{x}\right]\\ &=\frac{1}{2(e^2-5/2)}\left(\left[\frac{1}{2}e^{2x}\right]_0^2 - \Big[3(x-1)^3\Big]_1^2\right)\\ &=\frac{1}{2e^2-5}\left(\frac{1}{2}e^{4}-\frac{1}{2} -3\right)\\ &=\frac{e^4-7}{4e^2-10}\approx 2.4 \end{align*}2.3.3.24. (✳).
Solution.
2.3.3.25. (✳).
Solution.
- We use thin horizontal strips of width \(\dee{y}\) as in the figure above.
- When we rotate about the \(y\)-axis, each strip sweeps out a thin disk
- whose radius is \(r=6-y\) when \(4\le y\le 6\) (see the blue strip in the figure above), and whose radius is \(r=\sqrt{y}\) when \(0\le y\le 4\) (see the red strip in the figure above) and
- whose thickness is \(\dee{y}\) and hence
- whose volume is \(\pi r^2\,\dee{y} = \pi(6-y)^2\,\dee{y}\) when \(4\le y\le 6\) and whose volume is \(\pi r^2\,\dee{y} =\pi y\,\dee{y}\) when \(0\le y\le 4\text{.}\)
- As our bottommost strip is at \(y=0\) and our topmost strip is at \(y=6\text{,}\) the total volume is\begin{gather*} \pi \int_0^4 y\,\dee{y} + \pi \int_4^6 (6-y)^2\,\dee{y} \end{gather*}
2.3.3.26. (✳).
Solution.
- we use thin vertical strips of width \(\dee{x}\) as in the figure above.
- When we rotate about the line \(y=-1\text{,}\) each strip sweeps out a thin disk
- whose radius is \(r=T(x)-B(x)=e^x+1\) and
- whose thickness is \(\dee{x}\) and hence
- whose volume is \(\pi r^2\,\dee{x} = \pi(e^x+1)^2\,\dee{x}\text{.}\)
- As our leftmost strip is at \(x=0\) and our rightmost strip is at \(x=1\text{,}\) the total volume is\begin{align*} \pi \int_0^1 (e^x+1)^2\,\dee{x} &=\pi \int_0^1 (e^{2x}+2e^x+1)\,\dee{x}\\ \amp=\pi\bigg[\frac{e^{2x}}{2}+2e^x + x\bigg]_0^1\\ &=\pi\bigg[\Big(\frac{e^{2}}{2}+2e + 1\Big) -\Big(\frac{1}{2}+2 + 0\Big)\bigg]\\ &=\pi\Big(\frac{e^{2}}{2}+2e -\frac{3}{2}\Big) \end{align*}
2.3.3.27.
Solution.
2.3.3.28.
Solution.
To make things look a little cleaner, we use the substitution \(u=y-3\text{,}\) \(\dee{u}=\dee{y}\text{.}\) Then the limits of integration become \(-3\) and \(3\text{,}\) respectively, and \(y=u+3\text{.}\) (Geometrically, we’re re-centring the circle at the origin, instead of at the point (0,3).)
\begin{align*} &=\frac{\int_{-3}^3 (u+3)(2+u+3)\sqrt{9-u^2}\,\dee{u}}{\int_{-3}^3 \big(2+u+3\big)\sqrt{9-u^2}\,\dee{u}}\\ &=\frac{\int_{-3}^3 \big(u^2+8u+15\big)\sqrt{9-u^2}\,\dee{u}}{\int_{-3}^3 \big(u+5\big)\sqrt{9-u^2}\,\dee{u}} = \frac{N}{D}\tag{$*$}\\ \end{align*}Let’s start by finding \(D\text{,}\) the integral of the denominator. If we break it into two pieces, we can use symmetry and geometry to evaluate it.
\begin{align*} D&=\int_{-3}^3 u\sqrt{9-u^2}\,\dee{u}+5\int_{-3}^3 \sqrt{9-u^2}\,\dee{u}\\ \end{align*}The left integrand is odd, so its integral over a symmetric interval is 0. (You can also evaluate this using the substitution \(w=9-u^2\text{,}\) \(\dee{w}=-2u\,\dee{u}\text{.}\)) The right integral represents the area underneath half a circle of radius 3, centred at the origin.
\begin{align*} D&=0+5\cdot\frac{1}{2}\pi\cdot 3^2 = \frac{45}{2}\pi \end{align*}The first integrand is even, with a symmetric interval of integration, so we can simplify its limits of integration a little bit. The middle integrand is odd, so its integral over the symmetric interval \([-3,3]\) is zero. The last integral is the area of half a circle of radius 3.
\begin{align*} N&=2\int_{0}^3 u^2\sqrt{9-u^2}\,\dee{u}+0+15\cdot\pi\cdot3^2\\ &=\frac{135}{2}\pi+2\int_{0}^3 u^2\sqrt{9-u^2}\,\dee{u}\\ \end{align*}The remaining integral has a quadratic function underneath a square root with no obvious substitution, so we use a trigonometric substitution. Let \(u=3\sin\theta\text{,}\) \(\dee{u}=3\cos\theta\,\dee{\theta}\text{.}\) Note \(3\sin(0)=0\) and \(3\sin(\pi/2)=3\text{,}\) so the limits of integration become 0 and \(\frac{\pi}{2}\text{.}\)
\begin{align*} N&=\frac{135}{2}\pi+2\int_{0}^{\pi/2} \big(3\sin\theta\big)^2\sqrt{9-\big(3\sin\theta\big)^2}\cdot 3\cos\theta\,\dee{\theta}\\ &=\frac{135}{2}\pi+2\int_{0}^{\pi/2} 9\sin^2\theta \cdot \sqrt{9-9\sin^2\theta}\cdot 3\cos\theta\,\dee{\theta}\\ &=\frac{135}{2}\pi+54\int_{0}^{\pi/2} \sin^2\theta \cdot \sqrt{9\cos^2\theta}\cdot \cos\theta\,\dee{\theta}\\ &=\frac{135}{2}\pi+54\int_{0}^{\pi/2} \sin^2\theta \cdot 3\cos \theta\cdot \cos\theta\,\dee{\theta}\\ &=\frac{135}{2}\pi+162\int_{0}^{\pi/2} \sin^2\theta \cdot \cos^2 \theta\,\dee{\theta}\\ \end{align*}Using the identity \(\sin(2\theta)=2\sin\theta\cos\theta\text{,}\) we see \(\sin^2\theta\cos^2\theta =\big(\sin\theta\cos\theta\big)^2 = \frac{1}{4}\sin^2(2\theta)\)
\begin{align*} N&=\frac{135}{2}\pi+162\int_{0}^{\pi/2} \frac{1}{4}\sin^2(2\theta)\,\dee{\theta}\\ \end{align*}Now, we use the identity \(\sin^2 x = \frac{1}{2}(1-\cos(2x))\text{,}\) with \(x=2\theta.\)
\begin{align*} N&=\frac{135}{2}\pi+162\int_{0}^{\pi/2} \frac{1}{8}\big(1-\cos(4\theta)\big)\,\dee{\theta}\\ &=\frac{135}{2}\pi+\frac{81}{4}\int_{0}^{\pi/2} 1-\cos(4\theta)\,\dee{\theta}\\ &=\frac{135}{2}\pi+\frac{81}{4}\left[\theta-\frac{1}{4}\sin(4\theta)\right]_{0}^{\pi/2}\\ &=\frac{135}{2}\pi+\frac{81}{4}\left(\frac{\pi}{2}\right)\\ &=\frac{621}{8}\pi\\ \end{align*}Now, using equation (\(*\)), we find \(\bar y\text{:}\)
\begin{align*} \bar y &=\frac{N}{D} = \frac{\frac{621}{8}\pi}{\frac{45}{2}\pi} = \frac{69}{20}=3.45 \end{align*}2.3.3.29.
Solution.
-
To find the centre of mass of the rod \(R\text{,}\) we need to know its density at height \(y\text{,}\) \(\rho(y)\text{.}\) Since the mass of a section of \(R\) is the same as the volume of a section of the cone, let’s find the volume of a thin horizontal slice of the cone at height \(y\text{,}\) with thickness \(\dee{y}.\) To find its radius \(s\text{,}\) we use similar triangles. The diagram below represents a vertical cross-section of the cone.Since \(\frac{r}{h}=\frac{s}{h-y}\text{,}\) the radius of our slice at height \(y\) is \(s=\frac{r}{h}(h-y)\text{.}\) Then the volume of the slice is \(\pi s^2\dee{y}=\pi\left(\frac{r}{h}(h-y)\right)^2\,\dee{y}\text{.}\) Correspondingly, the mass of the piece of the rod at position \(y\) with length \(\dee{y}\) is \(\pi\left(\frac{r}{h}(h-y)\right)^2\,\dee{y}\text{,}\) so its density is\begin{equation*} \rho(y) = \frac{\pi\left(\frac{r}{h}(h-y)\right)^2\,\dee{y}}{\dee{y}}=\pi\left(\frac{r}{h}(h-y)\right)^2. \end{equation*}Now, we can find the centre of mass of \(R\text{:}\)\begin{align*} \bar y &= \frac{\int_0^h y\rho(y)\,\dee{y}}{\int_0^h \rho(y)\,\dee{y}} = \frac{\int_0^h y\pi\left(\frac{r}{h}(h-y)\right)^2\,\dee{y}}{\int_0^h \pi\left(\frac{r}{h}(h-y)\right)^2\,\dee{y}}\\ & = \frac{\frac{r^2}{h^2}\pi\int_0^h y\left(h-y\right)^2\,\dee{y}}{\frac{r^2}{h^2}\pi\int_0^h \left(h-y\right)^2\,\dee{y}}\\ & = \frac{\int_0^h \left(h^2y-2hy^2+y^3\right)\,\dee{y}}{\int_0^h \left(h^2-2hy+y^2\right)\,\dee{y}}\\ &=\frac{\left[\frac{h^2}{2}y^2-\frac{2h}{3}y^3+\frac{1}{4}y^4\right]_0^h} {\left[h^2y-hy^2+\frac{1}{3}y^3\right]_0^h}\\ &=\frac{\frac{h^4}{2}-\frac{2h^4}{3}+\frac{h^4}{4}} {h^3-h^3+\frac{1}{3}h^3}\\ &=\frac{h^4}{h^3}\cdot\frac{\frac{1}{2}-\frac{2}{3}+\frac{1}{4}}{\frac{1}{3}}=\frac{h}{4} \end{align*}So, the centre of mass of the cone occurs \(\frac{h}{4}\) metres above its base.Remark: it is quite interesting that the centre of mass does not depend on the radius of the cone!
- To find the centre of mass of a truncated cone, we simply consider a truncated rod. If the top \(h-k\) metres are missing, then the height of the cone (and also the rod) is \(k\text{.}\) Then the centre of mass has height:\begin{align*} \bar y &= \frac{\int_0^{k} y\rho(y)\,\dee{y}}{\int_0^h \rho(y)\,\dee{y}} = \frac{\int_0^{k} y\pi\left(\frac{r}{h}(h-y)\right)^2\,\dee{y}}{\int_0^h \pi\left(\frac{r}{h}(h-y)\right)^2\,\dee{y}}\\ & = \frac{\frac{r^2}{h^2}\pi\int_0^{k} y\left(h-y\right)^2\,\dee{y}}{\frac{r^2}{h^2}\pi\int_0^{k} \left(h-y\right)^2\,\dee{y}}\\ & = \frac{\int_0^{k} \left(h^2y-2hy^2+y^3\right)\,\dee{y}}{\int_0^{k} \left(h^2-2hy+y^2\right)\,\dee{y}}\\ &=\frac{\left[\frac{h^2}{2}y^2-\frac{2h}{3}y^3+\frac{1}{4}y^4\right]_0^{k}} {\left[h^2y-hy^2+\frac{1}{3}y^3\right]_0^{k}}\\ &=\frac{\frac{1}{2}h^2k^2 - \frac{2}{3}hk^3+\frac{1}{4}k^4}{h^2k-hk^2+\frac{1}{3}k^3}\\ &=\frac{\frac{1}{2}h^2k - \frac{2}{3}hk^2+\frac{1}{4}k^3}{h^2-hk+\frac{1}{3}k^2} \end{align*}
2.3.3.30.
Solution.
2.3.3.31.
Solution.
-
Option 1: As in Question 29, we’ll model the tank of water as a vertical rod, along the \(y\)-axis spanning the interval \([0,1]\text{,}\) such that the mass of a piece of the rod along \([a,b]\) is the same as the mass of the water from height \(y=a\) to height \(y=b\text{.}\) Then, the centre of mass of the rod will be the same as the centre of mass of the water.Consider a horizontal slice of water at height \(y\text{,}\) with thickness \(\dee{y}\text{.}\) If the radius of this slice is \(r(y)\text{,}\) then the volume of the slice is \(\pi r(y)^2\,\dee{y}\) m\(^3\text{,}\) so its mass is \(1000\pi r(y)^2\,\dee{y}\) kg. Then the mass of the slice of the rod at position \(y\) with length \(\dee{y}\) is \(1000\pi r(y)^2\,\dee{y}\) kg, so its density \(\rho(y)\) is\begin{equation*} \rho(y) = \frac{1000\pi r(y)^2\,\dee{y}\text{ kg}}{\dee{y}\text{ m}} = 1000\pi r(y)^2 \frac{\text{kg}}{\text{m}}. \end{equation*}So, let’s find \(r(y)\text{,}\) the radius of the slice of water at height \(y\text{.}\)Using the Pythagorean Theorem, \(r = \sqrt{1-y^2}\text{.}\) Therefore,\begin{equation*} \rho(y)= 1000\pi(1-y^2) \end{equation*}We use Equation 2.3.4 to calculate the centre of mass of the rod, which is the height of the centre of mass of Tank \(A\text{:}\)\begin{align*} \bar y_A&=\frac{\int_0^1 y\rho(y)\,\dee{y}}{\int_0^1 \rho(y)\,\dee{y}} = \frac{\int_0^1 1000\pi y(1-y^2)\,\dee{y}}{\int_0^1 1000\pi(1-y^2)\,\dee{y}}\\ &=\frac{\int_0^1(y-y^3)\,\dee{y}}{\int_0^1(1-y^2)\,\dee{y}} = \frac{\left[\frac{1}{2}y^2 - \frac{1}{4}y^4\right]_0^1}{\left[y-\frac{1}{3}y^3\right]_0^1} = \frac{\frac{1}{2}-\frac{1}{4}}{1-\frac{1}{3}}=\frac{3}{8} \text{m} \end{align*}From here, we can find the work done moving pumping the water to a height of 3 metres. We’ve moved the centre of mass from \(\bar y_A = \frac{3}{8}\) metres to 3 metres.\begin{align*} W&=\left(\frac{2000}{3}\pi \text{kg}\right)\times\left(3 - \frac{3}{8} \text{m}\right)\times\left(9.8 \frac{\text{m}}{\text{sec}^2}\right)\\ &=17,150\pi \text{J} \end{align*}
-
Option 2: We can use the techniques of Section 2.1 to calculate the amount of work it takes to pump the water from tank \(A\) to a height of 3 metres. That solves part (a), and we can use the amount of work to figure out the centre of gravity of the water in Tank \(A\) to help us solve part (b).At height \(y\text{,}\) a horizontal layer of water in Tank \(A\) forms a disk with thickness \(\dee{y}\) and radius \(\sqrt{1-y^2}\text{.}\) (The radius comes from the Pythagorean Theorem — see the diagram below.)The volume of the layer at height \(y\) is \(\pi \left(\sqrt{1-y^2}\right)^2\,\dee{y} = \pi(1-y^2)\,\dee{y}\) m\(^3\text{,}\) so its mass is \(1000\pi(1-y^2)\,\dee{y}\) kg.The layer at height \(y\) needs to be pumped a distance of \(3-y\) metres. So, the work involved pumping the layer at height \(y\) is:\begin{align*} \dee{W}&=\left(1000\pi(1-y^2)\,\dee{y} \text{kg}\right)\times\left(3-y \text{m}\right)\times\left(9.8 \text{m}/\text{sec}^2\right)\\ &=9800\pi(y^3-3y^2-y+3)\,\dee{y}\text{ J}\\ \end{align*}
Then the work involved pumping out the entire tank to a height of 3 metres is:
\begin{align*} W&=\int_0^1 9800\pi(y^3-3y^2-y+3)\,\dee{y}\\ &=9800\pi\left[\frac{1}{4}y^4 - y^3-\frac{1}{2}y^2+3y\right]_0^1\\ &=17,150\pi \text{J} \end{align*}This gives us an answer to part (a). To find the centre of mass of the water in Tank \(A\text{,}\) note that the work done is equivalent to moving a point mass from the centre of mass of the tank to a height of 3 metres. We know the water in Tank \(A\) has mass \(\frac{2000}{3}\pi\) kg. So, if \(\bar y_A\) is the centre of mass of the water in Tank \(A\text{:}\)\begin{align*} W&=\left(\frac{2000}{3}\pi \text{kg}\right)\times\left(3-\bar y_A \text{m}\right)\times\left(9.8 \text{m}/\text{sec}^2\right)\\ 17,150\pi&=\left(\frac{2000}{3}\pi\right)\left(3-\bar y_A\right)(9.8)\\ \frac{21}{8}&=3-\bar y_A\\ \bar y_A &= \frac{3}{8} \text{m} \end{align*}
As a percentage of 17,150\(\pi\text{,}\) this is:
\begin{align*} \text{waste}&= \left(\frac{19,600\pi\left(1-\frac{\pi}{9}\right)}{17,150\pi}\right)\times 100\\ &=\frac{8}{7}\left(1-\frac{\pi}{9}\right)\times 100\approx 74\% \end{align*}2.3.3.32.
Solution.
3
2.4 Separable Differential Equations
2.4.7 Exercises
2.4.7.1.
Solution.
- If \(y=5(e^x-3x^2-6x-6)\text{,}\) then \(\diff{y}{x} = 5(e^x-6x-6)\text{.}\) Let’s see whether this is equal to \(y+15x^2\text{:}\)\begin{align*} y+15x^2&=5(e^x-3x^2-6x-6)+15x^2\\ &=5(e^x-3x^2-6x-6+3x^2)\\ &=5(e^x-6x-6)\\ &=\diff{y}{x} \end{align*}So, \(y=5(e^x-3x^2-6x-6)\) is indeed a solution to the differential equation \(\diff{y}{x}=y+15x^2\text{.}\)
- If \(y=\dfrac{-2}{x^2+1}\text{,}\) then \(\diff{y}{x} = \dfrac{4x}{(x+1)^2}\text{.}\) Let’s see whether this is equal to \(xy^2\text{:}\)\begin{align*} xy^2&=x\left(\frac{-2}{x^2+1}\right)^2\\ &=\frac{4x}{(x^2+1)^2}\\ &=\diff{y}{x} \end{align*}So, \(y=\dfrac{-2}{x^2+1}\) is indeed a solution to the differential equation \(\diff{y}{x}=xy^2\text{.}\)
- If \(y=x^{3/2}+x\text{,}\) then \(\diff{y}{x} = \frac{3}{2}\sqrt{x}+1\text{.}\)\begin{align*} \left(\diff{y}{x}\right)^2+\diff{y}{x}&=\left(\frac{3}{2}\sqrt{x}+1\right)^2+\frac{3}{2}\sqrt{x}+1\\ &=\frac{9}{4}x+\frac{9}{2}\sqrt{x}+2\\ &\neq \diff{y}{x} \end{align*}So, \(y=x^{3/2}+x\) is not a solution to the differential equation \(\left(\diff{y}{x}\right)^2+\diff{y}{x}=y\text{.}\)
2.4.7.2.
Solution.
- \(3y\diff{y}{x}=x\sin y\) can be written as \(\diff{y}{x} = x\left(\frac{\sin y}{3y}\right)\text{,}\) which fits the form of a separable equation with \(f(x)=x\text{,}\) \(g(y) = \frac{\sin y}{3y}\text{.}\)
- \(\diff{y}{x} = e^{x+y} = e^xe^y\) which fits the form of a separable equation using \(f(x) = e^x\text{,}\) \(g(y) = e^y\text{.}\)
- \(\diff{y}{x}+1=x\) can be written as \(\diff{y}{x} = (x-1)\text{,}\) which fits the form of a separable equation using \(f(x)=x-1\text{,}\) \(g(y)=1\text{.}\) (We can solve it by simply antidifferentiating.)
- Notice the left side of the equation \(\left(\diff{y}{x}\right)^2-2x\diff{y}{x}+x^2=0\) is a perfect square. So, this equation is equivalent to \(\left(\diff{y}{x}-x\right)^2=0\text{,}\) that is, \(\diff{y}{x}=x\text{.}\) This has the form of a separable equation with \(f(x)=x\text{,}\) \(g(y)=1\text{.}\)
2.4.7.3.
Solution.
2.4.7.4.
Solution.
Since \(y=f(x)\) is a solution, we know \(\diff{}{x}\{f(x)\}=xf(x)\text{.}\) Also, \(\diff{}{x}\{f(x)+C\}=\diff{}{x}\{f(x)\}\text{.}\) So, \(\diff{}{x}\{f(x)+C\}=xf(x)\text{.}\)
\begin{align*} xf(x)&=x(f(x)+C)\\ 0&=xC \end{align*}2.4.7.5.
Solution.
- Since \(|y| \geq 0\) no matter what \(y\) is, we see \(Cx \ge 0\) for all \(x\) in the domain of \(f(x)\text{.}\) Since \(C\) is positive, that means the domain of \(f(x)\) only includes nonnegative numbers. So, the largest possible domain of \(f(x)\) is \([0,\infty)\text{.}\)
-
None exists.The graph of \(Cx\) is given below for some positive constant \(C\text{,}\) also with the graph of \(-Cx\text{.}\) If \(y=f(x)\) were sometimes the top function, and other times the bottom function, then there would be a jump discontinuity where it switched. Then the derivative of \(f(x)\) would not exist, violating the second property.A tiny technical note is that it’s possible that \(f(x)=Cx\) when \(x=0\) and \(f(x)=-Cx\) when \(x \gt 0\) (or vice-versa). This would not introduce a jump discontinuity, but it also does not satisfy that \(f(x) \gt 0\) for some values of \(x\text{.}\)
2.4.7.6.
Solution.
2.4.7.7.
Solution.
2.4.7.8.
Solution.
- When \(y=0\text{,}\) \(y'=\frac{0}{2}-1=-1\text{.}\)
- When \(y=2\text{,}\) \(y'=\frac{2}{2}-1=0\text{.}\)
- When \(y=3\text{,}\) \(y'=\frac{3}{2}-1=0.5\text{.}\)
-
The small red lines have varying slopes. The red lines on points with \(y\)-coordinate 2 have slopes of \(0\text{;}\) this matches \(y'\) when \(y=0\text{,}\) as we saw above. The red lines on points with \(y\)-coordinate 0 have slopes of approximately \(-1\text{;}\) again, this matches what we found for \(y'\) when \(y=0\text{.}\)The red lines correspond to a tiny section of \(y(x)\text{,}\) if \(y(x)\) passes through that point. So, we can sketch a possible curve \(y(x)\) satisfying the equation by starting somewhere, then following the slopes.For example, suppose we start at the origin.Then our function is decreasing at that point, which leads us to a coordinate where (as we see from the red marks) the function is decreasing slightly faster.Following the red marks leads us down even further, so our function \(y(x)\) might look something like this:However, we didn’t have to start at the origin. Suppose \(y(0)=3\text{.}\) Then at \(x=0\text{,}\) \(y\) is increasing, with slope \(\frac{1}{2}\text{.}\)Our red marks run out that high up, but we now \(y'=\frac{1}{2}y-1\text{,}\) so \(y'\) increases as \(y\) increases. That means our function keeps getting steeper and steeper, possibly something like this:If \(y(0)=2\text{,}\) we see another possible curve is the constant function \(y(x)=2\text{.}\)
2.4.7.9.
Solution.
- If \(y(1)=0\text{,}\) then \(y'(1)=0-\frac{1}{2}=-\frac{1}{2}\text{.}\)
- If \(y(1)=2\text{,}\) then \(y'(1)=2-\frac{1}{2}=\frac{3}{2}\text{.}\)
- If \(y(1)=-2\text{,}\) then \(y'(1)=-2-\frac{1}{2}=-\frac{5}{2}\text{.}\)
-
There are \(7 \times 7 = 49\) points on the grid; we don’t want to make 49 separate calculations. Let’s find some shortcuts.
- If \(y'(x)=0\text{,}\) then \(y=\frac{x}{2}\text{,}\) which applies to the points \((0,0)\text{,}\) \((2,1)\text{,}\) \((4,2)\) and \((6,3)\text{.}\) These are the orange dots in the sketch below.
- If \(y'(x)=1\text{,}\) then \(y=1+\frac{x}{2}\text{,}\) which applies to the points \((0,1)\text{,}\) \((2,2)\text{,}\) and \((4,3)\text{.}\) (Note these are exactly 1 unit above the points with \(y'=0\text{.}\)) These are the red dots in the sketch below.
- If \(y'(x)=-1\text{,}\) then \(y=-1+\frac{x}{2}\text{,}\) which applies to the points \((0,-1)\text{,}\) \((2,-2)\text{,}\) and \((4,-3)\text{.}\) (Note these are exactly 1 unit below the points with \(y'=0\text{.}\)) These are the yellow dots in the sketch below.
- If \(x\) increases and \(y\) stays the same, \(y\) decreases.
- If \(y\) increases and \(x\) stays the same, \(y\) increases.
- If we draw a straight line of slope \(\dfrac{1}{2}\) on our sketch, for every point on that line, our mark has the same slope: for instance, the points where we draw a mark with slope 0 are \((0,0)\text{,}\) \((2,1)\text{,}\) and \((4,2)\text{,}\) and these all lie on the line \(f(x)=\frac{x}{2}\text{.}\)
This is enough to give us a pretty good sketch. The points whose slopes we found explicitly have dots; the rest can be sketched as either steeper or less steep than what’s near them. -
To sketch a possible graph of \(y(x)\text{,}\) we choose a point \((x,y(x))\text{,}\) then follow the red lines.For example, if we suppose that \(y(4)=2\text{,}\) then near \((4,2),\) the lines tell us \(y(x)\) is fairly flat; and it is increasing to the left of \(x=4\text{,}\) and decreasing to the right.Following the red lines a little farther in each direction brings us somewhere like this:Extending yet further, we might sketch something like the following:By choosing another point \((x,y(x))\) to be on the curve, we might find other potential curves. Some examples are shown below.
2.4.7.10. (✳).
Solution.
Integrating both sides:
\begin{align*} \int e^y\,\dee{y} &= \int 2x\,\dee{x}\\ e^y &= x^2+C\\ \end{align*}Since \(y=\log2\) when \(x=0\text{,}\) we have
\begin{align*} e^{\log 2} &= 0^2+C\\ 2 &= C,\\ \end{align*}and therefore
\begin{align*} e^{y} &= x^2+2\\ y &= \log(x^2+2) \end{align*}2.4.7.11. (✳).
Solution.
To satisfy \(y(0)=3\text{,}\) we need \(\log 3 = \frac{1}{2}\log(1+0)+C\text{,}\) so \(C=\log 3\text{.}\) Thus:
\begin{align*} \log |y| &= \frac{1}{2}\log(1+x^2)+\log 3\\ &=\log\sqrt{1+x^2}+\log 3\\ &=\log 3\sqrt{1+x^2}\\ \end{align*}So,
\begin{align*} |y|&=3\sqrt{1+x^2}\\ \end{align*}We are told to find a function \(y(x)\text{.}\) So far, we have two possible functions from the work above: maybe \(y=3\sqrt{1+x^2}\text{,}\) and maybe \(y=-3\sqrt{1+x^2}\text{.}\) It’s important to note that \(y=\pm3\sqrt{1+x^2}\) is not a function: for an equation to represent a function, for every input in the domain, there must only be one output. That is, functions pass the vertical line test. (See the CLP-1 text for a definition of the vertical line test and a formal definition of a function.) So, we need to decide whether our function is \(y=3\sqrt{1+x^2}\) or \(y=-3\sqrt{1+x^2}\text{.}\) Since \(y(0)=3\text{,}\) we conclude
\begin{align*} y(x)&=3\sqrt{1+x^2} \end{align*}2.4.7.12. (✳).
Solution.
2.4.7.13. (✳).
Solution.
We can guess the antiderivative of \(xe^{x^2}\text{,}\) or use the substitution \(u=x^2\text{,}\) \(\dee{u}=2x\dee{x}\text{.}\)
\begin{align*} \frac{y^3}{3}&=\frac{1}{2} e^{x^2}+C'\\ y^3&=\frac{3}{2} e^{x^2}+3C'\\ \end{align*}Since \(C'\) can be any constant in \((-\infty,\infty)\text{,}\) then also \(3C'\) can be any constant in \((-\infty,\infty)\text{,}\) so we replace \(3C'\) with the arbitrary constant \(C\text{.}\)
\begin{align*} y^3&=\frac{3}{2} e^{x^2}+C\\ y&=\root{3}\of{\frac{3}{2} e^{x^2}+C} \end{align*}2.4.7.14. (✳).
Solution.
Since \(C\) can be any constant in \((-\infty,\infty)\text{,}\) then also \(-C\) can be any constant in \((-\infty,\infty)\text{,}\) so we write \(C\) instead of \(-C\text{.}\)
\begin{align*} e^{-y}&=C-\frac{1}{2} x^2\\ -y&=\log\left(C-\frac{x^2}{2}\right)\\ y&=-\log\left(C-\frac{x^2}{2}\right) \end{align*}2.4.7.15. (✳).
Solution.
Integrating both sides, we find
\begin{align*} \int y^2 \,\dee{y} &=\int (e^x - 2x ) \,\dee{x}\\ \frac{1}{3}y^3 &= e^x - x^2 + C\\ \end{align*}Setting \(x = 0\) and \(y = 3\text{,}\) we find \(\frac{1}{3}3^3=e^0-0^2+C\) and hence \(C=8\text{.}\)
\begin{align*} \frac{1}{3}y^3 &= e^x - x^2 + 8\\ y &= (3e^x -3x^2+ 24)^{1/3} \end{align*}2.4.7.16. (✳).
Solution.
To have \(y=-\frac{1}{4}\) when \(x=0\text{,}\) we must choose \(C\) to obey
\begin{align*} {\Big(-\frac{1}{4} \Big)}^{-2}& =0 +2C\\ 16&= 2C\\ \end{align*}So, from (\(*\)),
\begin{align*} y^{-2}&=x^2+2C =x^2+16\\ y^2&=\frac{1}{x^2+16}\\ \end{align*}Now, we have two potential candidates for \(y(x)\text{:}\)
\begin{align*} y&=\frac{1}{\sqrt{x^2+16}}\qquad\text{OR}\qquad y=-\frac{1}{\sqrt{x^2+16}}\\ \end{align*}We know \(y=-\frac14\) when \(x=0\text{.}\) The only function above that fits this is
\begin{align*} y&=-\frac{1}{\sqrt{x^2+16}} \end{align*}2.4.7.17. (✳).
Solution.
Plugging in \(x=1\) and \(y=4\) gives \(\frac{4^2}{2} = 5+2+3+C,\) and so \(C=-2\text{.}\) Therefore
\begin{align*} \frac{y^2}{2} &= 5x^3 + 2x^2 + 3x - 2\\ y^2 &= 10x^3 + 4x^2 + 6x - 4\\ \end{align*}This leaves us with two possible functions for \(y\text{:}\)
\begin{align*} y = \sqrt{10x^3 + 4x^2 + 6x - 4} &\qquad\text{or}\qquad y = - \sqrt{10x^3 + 4x^2 + 6x - 4}\\ \end{align*}When \(x=1\text{,}\) \(y=4\text{.}\) This only fits the first equation, so
\begin{align*} y &= \sqrt{10x^3 + 4x^2 + 6x - 4} \end{align*}2.4.7.18. (✳).
Solution.
We are told that \(y=1\) when \(x=0\text{.}\) That is, \(1=e^0e^C\text{,}\) so \(e^C=1\text{.}\) That is, \(C=0\text{.}\)
\begin{align*} |y|&=e^{x^4/4}\\ \end{align*}This leaves us with two potential functions:
\begin{align*} y=e^{x^4/4}&\qquad\text{or}\qquad y=-e^{x^4/4}\\ \end{align*}The first is always positive, and the second is always negative. Since \(y=1\) (a positive number) when \(x=0\text{,}\) we see
\begin{align*} y&=e^{x^4/4} \end{align*}2.4.7.19. (✳).
Solution.
Using the method of partial fractions, we see \(\frac{1}{y(y-1)} = \frac{1}{y-1}-\frac{1}{y}.\)
\begin{align*} \left(\frac{1}{y-1}-\frac{1}{y}\right)\,\dee{y} &=\frac{\dee{x}}{x}\\ \int\left(\frac{1}{y-1}-\frac{1}{y}\right)\,\dee{y} &=\int\frac{\dee{x}}{x}\\ \log|y-1|-\log|y| &=\log |x| + C\\ \log\frac{|y-1|}{|y|}&=\log|x|+C \tag{$*$}\\ \end{align*}To determine \(C\) we set \(x=1\) and \(y=-1\text{.}\)
\begin{align*} \log\frac{|-2|}{|-1|}&=\log|1|+C\\ \log 2&=C\\ \end{align*}Returning to (\(*\)),
\begin{align*} \log\frac{|y-1|}{|y|}&=\log|x|+\log 2\\ \log\left|\frac{y-1}{y}\right|&=\log|2x|\\ \left|\frac{y-1}{y}\right|&=|2x|\\ \end{align*}As \(y(1)=-1\) is an initial condition, we have that \(x\ge 1\) and \(|2x|=2x\text{.}\) For \(x=1\text{,}\) we have \(y=-1\text{.}\) So at least for \(x\) near \(1\text{,}\) we have \(y\) near \(-1\text{,}\) so that \(\frac{y-1}{y}\) is positive and we may drop the absolute value signs. There remains the possibility that \(\frac{y(x)-1}{y(x)}\) changes sign for some larger \(x\gt 1\text{.}\) For now, we will simply ignore that possibility. At the end, we will explicitly check that the \(y(x)\) we come up with really does satisfy the differential equation \(x\diff{y}{x}+y=y^2\) and the initial condition \(y(1)=-1\text{.}\)
\begin{align*} \frac{y-1}{y}&=2x\\ y-1&=2xy\\ y-2xy&=1\\ y(1-2x)&=1\\ y&=\frac{1}{1-2x} \end{align*}So, our differential equation is satisfied. Furthermore:
\begin{align*} y(1)&=\frac{1}{1-2\times 1}=-1 \end{align*}2.4.7.20. (✳).
Solution.
To determine \(C\) we set \(x=0\) and \(y=e\text{.}\)
\begin{align*} \log e &=\frac{0^2}{2}+ C\\ 1&= C\\ \end{align*}So, the solution is
\begin{align*} \log|y|&= \frac{x^2}{2} + 1\\ \end{align*}We are told that \(y=f(x) \gt 0\text{,}\) so may drop the absolute value signs.
\begin{align*} \log y &= \frac{x^2}{2}+1\\ y&=e^{1+\frac{1}{2}x^2}=e\cdot e^{x^2/2} \end{align*}2.4.7.21. (✳).
Solution.
Using partial fractions decomposition, we find \(\frac{1}{x(x+1)} = \frac{1}{x}-\frac{1}{x+1}\text{.}\)
\begin{align*} y\,\dee{y} &=\left(\frac{1}{x}-\frac{1}{x+1}\right)\,\dee{x}\\ \int y\,\dee{y} &=\int\left(\frac{1}{x}-\frac{1}{x+1}\right)\,\dee{x}\\ \frac{y^2}{2}&=\log|x|-\log|x+1|+C = \log\left| \frac{x}{x+1}\right|+C\\ \end{align*}To satisfy the initial condition \(y(1)=2\) we must choose \(C\) to obey
\begin{align*} \frac{2^2}{2} &= \log\left|\frac{1}{1+1}\right|+C\\ 2&=\log\frac12+C\\ C&=2-\log\frac{1}{2}\\ \end{align*}So,
\begin{align*} \frac{y^2}{2}&= \log\left| \frac{x}{x+1}\right|+2-\log\frac{1}{2}\\ y^2&= 2\log\left| \frac{x}{x+1}\right|+4-2\log\frac{1}{2}\\ \end{align*}Note that the question specifies that \(y(1)=2\) is an initial condition. So we always have \(x\ge 1\text{.}\) Then \(\frac{x}{x+1}\) is positive, and we can drop the absolute values.
\begin{align*} y^2&= 2\log \frac{x}{x+1}+4-2\log\frac{1}{2}\\ \end{align*}This leaves two options for \(y(x)\text{:}\) the positive or negative square root of the right hand side above. Since \(y(1)=1\text{,}\) which is positive, we must choose the positive square root.
\begin{align*} y(x)&=\sqrt{2\Big(\log\frac{x}{x+1}-\log\frac{1}{2}+2\Big)}\\ &=\sqrt{4+2\log\frac{2x}{x+1}} \end{align*}2.4.7.22. (✳).
Solution.
For the integral on the left, we use the substitution \(u={y^2-4}\text{,}\) \(\frac12\dee{u}=y\,\dee{y}\text{.}\)
\begin{align*} \frac{1}{2}\int \big(1+\sqrt{u}\big)\,\dee{u}&=\sec x +C\\ \frac{1}{2} \left(u+\frac{2}{3}u^{3/2}\right)&=\sec x +C\\ \frac{1}{2} \left(y^2-4+\frac{2}{3}(y^2-4)^{3/2}\right)&=\sec x +C\\ y^2+\frac{2}{3}(y^2-4)^{3/2}&=2\sec x +2C+4\\ \end{align*}To find \(C\) we set \(x=0\) and \(y=2\text{.}\)
\begin{align*} 4+\frac{2}{3}\sqrt{4-4}^3&=2\sec (0)+2C+4\\ 4&=2+2C+4\\ 2&=2C+4\\ \end{align*}So,
\begin{align*} y^2+\frac{2}{3}(y^2-4)^{3/2}&=2\sec x +2 \end{align*}2.4.7.23. (✳).
Solution.
At \(t=0\text{,}\) \(P=90,000\) so
\begin{align*} 2\sqrt{90,000}&=-k\times 0+C\\ C&=2\times 300=600\\ \end{align*}Therefore,
\begin{align*} 2\sqrt{P}&=-kt+600\tag{$*$}\\ \end{align*}Now, we find \(k\text{.}\) Let \(t\) be measured in weeks. Then when \(t=6\text{,}\) \(P=40,000\text{.}\)
\begin{align*} 2\sqrt{40,000}&=-6k+600\\ 2\cdot 200&=-6k+600\\ k&=\frac{200}{6} = \frac{100}{3}\\ \end{align*}Substituting our value of \(k\) into (\(*\)):
\begin{align*} 2\sqrt{P}&=-\frac{100}{3}t+600\\ \end{align*}To find when the population will be 10,000, we set \(P=10,000\) and solve for \(t\text{.}\)
\begin{align*} 2\sqrt{10,000}&=-\frac{100}{3}t+600\\ 2\cdot 100&=-\frac{100}{3}t+600\\ \frac{100}{3}t&=400\\ t&=12 \end{align*}2.4.7.24. (✳).
Solution.
The left integral looks something like the antiderivative of arctangent. Let’s factor out that \(mg\) from the denominator.
\begin{align*} \frac{1}{mg}\int\frac{m}{1+\frac{k}{mg}v^2}\,\dee{v}&=-t+C\\ \frac{1}{g}\int\frac{1}{1+\left(\sqrt{\frac{k}{mg}}v\right)^2}\,\dee{v}&=-t+C\\ \end{align*}Now it looks even more like the derivative of arctangent. We can guess the antiderivative from here, or use the substitution \(u=\sqrt{\frac{k}{mg}}v\text{,}\) \(\dee{u}=\sqrt{\frac{k}{mg}}\,\dee{v}\text{.}\)
\begin{align*} \frac{1}{g}\sqrt{\frac{mg}{k}}\arctan\left(\sqrt{\frac{k}{mg}}v\right)&=-t+C\\ \sqrt{\frac{m}{gk}}\arctan\left(\sqrt{\frac{k}{mg}}v\right)&=-t+C\tag{$*$}\\ \end{align*}At \(t=0\text{,}\) \(v=v_0\text{,}\) so:
\begin{align*} \sqrt{\frac{m}{gk}}\arctan\left(\sqrt{\frac{k}{mg}}v_0\right)&=C\\ \end{align*}Plug \(C\) into \((*)\text{.}\)
\begin{align*} \sqrt{\frac{m}{gk}}\arctan\left(\sqrt{\frac{k}{mg}}v\right)&=\sqrt{\frac{m}{gk}}\arctan\left(\sqrt{\frac{k}{mg}}v_0\right)-t\\ \end{align*}At its highest point, the object has velocity \(v=0\text{.}\) This happens when \(t\) obeys:
\begin{align*} \sqrt{\frac{m}{gk}}\arctan\left(\sqrt{\frac{k}{mg}}0\right)&=\sqrt{\frac{m}{gk}}\arctan\left(\sqrt{\frac{k}{mg}}v_0\right)-t\\ 0&=\sqrt{\frac{m}{gk}}\arctan\left(\sqrt{\frac{k}{mg}}v_0\right)-t\\ t&=\sqrt{\frac{m}{gk}}\arctan\left(\sqrt{\frac{k}{mg}}v_0\right) \end{align*}2.4.7.25. (✳).
Solution.
At \(t=0\text{,}\) \(v=40\) so
\begin{align*} \frac{1}{40}&=k\times 0+ C\\ C&=\frac{1}{40}\\ \end{align*}Therefore,
\begin{align*} v(t)&=\frac{1}{kt+C}=\frac{1}{kt+1/40}=\frac{40}{40kt+1}\tag{$*$}\\ \end{align*}The constant of proportionality \(k\) is determined by
\begin{align*} v(10)&=20\\ 20&=\frac{40}{40k\times 10+1}\\ \frac{1}{2}&=\frac{1}{400k+1}\\ 400k+1&=2\\ k&=\frac{1}{400} \end{align*}We want to know the value of \(t\) that gives \(v(t)=5\text{.}\)
\begin{align*} 5&=\frac{40}{t/10+1}\\ \frac{t}{10}+1&=8\\ t&=70\text{ sec} \end{align*}2.4.7.26. (✳).
Solution.
Using the method of partial fractions, we find \(\frac{1}{(x-2)(x-3)}=\frac{1}{x-3}-\frac{1}{x-2}\text{.}\)
\begin{align*} \int\Big[\frac{1}{x-3}-\frac{1}{x-2}\Big]\dee{x}&= \int k\dee{t}\\ \log|x-3|-\log|x-2|&=kt+C\\ \log\left|\frac{x-3}{x-2}\right|&=kt+C\\ \left|\frac{x-3}{x-2}\right|&=e^{kt+C}=e^{kt}e^C\\ \frac{x-3}{x-2}&=De^{kt}\\ \end{align*}where \(D=\pm e^C\text{.}\) When \(t=0\text{,}\) \(x=1\text{,}\) forcing
\begin{align*} \frac{1-3}{1-2}&=De^{0}\\ D&=2\\ \end{align*}Hence
\begin{align*} \frac{x-3}{x-2}&=2e^{kt}\\ x-3&=2e^{kt}(x-2)\\ x-2e^{kt}x&=3-4e^{kt}\\ x(t)&=\frac{3-4e^{kt}}{1-2e^{kt}} \end{align*}2.4.7.27. (✳).
Solution.
Using the method of partial fractions, we see \(\frac{1}{P(4-P)}=\frac{1/4}{P} + \frac{1/4}{4-P}\text{.}\)
\begin{align*} \frac{1}{4}\Big[\frac{1}{P}+\frac{1}{4-P}\Big]\dee{P}&=\dee{t}\\ \int \frac{1}{4}\Big[\frac{1}{P}+\frac{1}{4-P}\Big]\dee{P}&=\int\dee{t}\\ \frac{1}{4}\big[\log|P|-\log|4-P|\Big]&=t+C\\ \end{align*}When \(t=0\text{,}\) \(P=2\text{,}\) so \(\frac{1}{4}\big[\log|2|-\log|2|\big]=C\implies C=0\text{.}\) So,
\begin{align*} \frac{1}{4}\log\Big|\frac{P}{4-P}\Big|&=t\\ \end{align*}At time \(t=0\text{,}\) \(\frac{P}{4-P}=1 \gt 0\text{.}\) The ratio may not change sign at any finite time, because this could only happen if at some finite time \(P\) took either the value 0 or the value 4. But at this time \(t=\frac{1}{4}\log\big|\frac{P}{4-P}\big|\) would have to be infinite. So \(\frac{P}{4-P} \gt 0\) for all time and:
\begin{align*} \frac{1}{4}\log \frac{P}{4-P} &=t\\ \log \frac{P}{4-P} &=4t\\ \frac{P}{4-P} &=e^{4t}\\ P&=(4-P)e^{4t}\\ P+Pe^{4t}&=4e^{4t}\\ P&=\frac{4e^{4t}}{1+e^{4t}}=\frac{4}{1+e^{-4t}} \end{align*}2.4.7.28. (✳).
Solution.
- The rate of change of speed at time \(t\) is \(-kv(t)^2\) for some constant of proportionality \(k\) (to be determined--but we assume it is positive, since the speed is decreasing). So \(v(t)\) obeys the differential equation \(\diff{v}{t}=-kv^2\) .
- The equation \(\diff{v}{t}=-kv^2\) is a separable differential equation, which we can solve in the usual way.\begin{align*} \diff{v}{t}&=-kv^2\\ \frac{\dee{v}}{-v^2}&=k\dee{t}\\ \int -\frac{\dee{v}}{v^2} &=\int k\dee{t}\\ \frac{1}{v}&=kt+C\\ \end{align*}
At time \(t=0\text{,}\) \(v=400\text{,}\) so \(C=\frac{1}{400}\text{.}\) Then:
\begin{align*} \frac{1}{v}&=kt+\frac{1}{400}\tag{$*$}\\ \end{align*}At time \(t=1\text{,}\) \(v=200\text{,}\) so
\begin{align*} \frac{1}{200}&=k+\frac{1}{400}\\ k&=\frac{1}{400}\\ \end{align*}Therefore, from (\(*\)),
\begin{align*} \frac{1}{v}&=\frac{t}{400}+\frac{1}{400} = \frac{t+1}{400}\\ v&=\frac{400}{t+1} \end{align*} - To find when the speed is 50, we set \(v=50\) in the equation from (b) and solve for \(t\text{.}\)\begin{align*} 50&=\frac{400}{t+1}\\ 50(t+1)&=400\\ t+1&=8\\ t&=7 \end{align*}
2.4.7.29. (✳).
Solution.
Since \(B(t)\) is our bank account balance and we’re not withdrawing money, \(B(t)\) is positive, so we can drop the absolute value signs.
\begin{align*} \log B(t)&= 0.06 t-0.02\cos t +C'\\ B(t)&= e^{0.06 t-0.02\cos t}e^{C'}\\ B(t)&=Ce^{0.06 t-0.02\cos t} \end{align*}So, when \(t=2\text{,}\)
\begin{align*} B(2)&=\underbrace{\vphantom{p}1000e^{0.02}}_{C} e^{0.06\times 2-0.02\cos 2}=\$1159.89 \end{align*}2.4.7.30. (✳).
Solution.
2.4.7.31. (✳).
Solution.
Using the method of partial fractions, we see \(\frac{1}{(y-2)(y-1)}=\frac{1}{y-2}-\frac{1}{y-1}\text{.}\)
\begin{align*} \int\Big[\frac{1}{y-2}-\frac{1}{y-1}\Big]\dee{y}&=\int\sin x\dee{x}\\ \log|y-2|-\log|y-1|&=-\cos x+c\\ \log\left|\frac{y-2}{y-1}\right|&=-\cos x+c\\ \left|\frac{y-2}{y-1}\right|&= e^{c-\cos x}\\ \end{align*}The condition \(y(0)=3\) forces \(\big|\frac{3-2}{3-1}\big|= e^{c-1}\) or \(e^c=\frac{1}{2} e\text{,}\) hence
\begin{align*} \left|\frac{y-2}{y-1}\right|&=\frac{1}{2} e^{1-\cos x} \end{align*}Solving for \(y\text{,}\)
\begin{align*} \frac{y-2}{y-1}&=\frac{1}{2} e^{1-\cos x}\\ 2(y-2)&=e^{1-\cos x}(y-1)\\ 2y-4&=ye^{1-\cos x}-e^{1-\cos x}\\ y\big(2-e^{1-\cos x}\big)&=4-e^{1-\cos x}\\ y&=\frac{4-e^{1-\cos x}}{2-e^{1-\cos x}} \end{align*}2.4.7.32. (✳).
Solution.
At time \(0\text{,}\) the height is \(6\text{,}\) so \(C=2\sqrt{6}\) and
\begin{align*} 2\sqrt{h}&=-{\Big(\frac{0.01}{3}\Big)}^{\!2} \sqrt{2g}\,t+2\sqrt{6}\\ \end{align*}We want to know when the height of the water in the tank is 0.
\begin{align*} 0&=-\Big(\frac{0.01}{3}\Big)^{\!2} \sqrt{2g}\,t+2\sqrt{6}\\ \Big(\frac{0.01}{3}\Big)^{\!2} \sqrt{2g}\,t&=2\sqrt{6}\\ t&=\frac{2\sqrt6}{\Big(\frac{0.01}{3}\Big)^{\!2} \sqrt{2g}}\\ &=2{\Big(\frac{3}{0.01}\Big)}^{\!2} \sqrt{\frac{3}{g}}\\ &=180,000 \sqrt{\frac{3}{g}}\approx 99,591\text{ sec} \approx 27.66\text{ hr} \end{align*}2.4.7.33. (✳).
Solution.
At time \(0\text{,}\) the height is \(12\text{,}\) so \(C=\frac{12^{5/2}}{5/2}-12\frac{12^{3/2}}{3/2} =12^{5/2}\big(\frac{2}{5}-\frac{2}{3}\big) =-\frac{4}{15}12^{5/2}\text{,}\) which yields
\begin{align*} \frac{h^{5/2}}{5/2}-12\frac{h^{3/2}}{3/2}&=\frac{1}{144} \sqrt{2g}\,t -\frac{4}{15}12^{5/2}\\ \end{align*}We want to find the time \(t\) when the height is \(h=0\text{.}\)
\begin{align*} 0&=\frac{1}{144} \sqrt{2g}\,t-\frac{4}{15}12^{5/2}\\ \frac{1}{144} \sqrt{2g}\,t&=\frac{4}{15}12^{5/2}\\ t&=\frac{4\times 144}{15} \sqrt{\frac{12^5}{2g}}\\ &=38.4 \sqrt{\frac{124416}{g}}\approx 2,394\,\text{sec }\approx 0.665\, \text{hr} \end{align*}2.4.7.34. (✳).
Solution.
Using the method of partial fractions,
\begin{align*} \int\left(\frac{1}{y-2}-\frac{1}{y-1}\right)\dee{y}&=\int \dee{x}\\ \log|y-2|-\log|y-1|&=x+C\\ \log\left|\frac{y-2}{y-1}\right|&=x+C\\ \end{align*}Observe that \(\diff{y}{x}=(y-1)(y-2) \gt 0\) for all \(y\ge 2\text{.}\) That is, \(f(x)\) is increasing at all \(x\) for which \(f(x) \gt 2\text{.}\) As \(f(0)=3\text{,}\) \(f(x)\) increases for all \(x\ge 0\text{,}\) and \(f(x)\ge 3\) for all \(x\ge 0\text{.}\) So we may drop the absolute value signs.
\begin{align*} \log\frac{f(x)-2}{f(x)-1}&=x+C\\ \frac{f(x)-2}{f(x)-1}&=e^Ce^x\\ \end{align*}At \(x=0\text{,}\) \(\frac{f(x)-2}{f(x)-1}=\frac{1}{2}\) so \(e^C=\frac{1}{2}\text{.}\)
\begin{align*} \frac{f(x)-2}{f(x)-1}&=\frac{1}{2} e^x\\ 2f(x)-4&=[f(x)-1]e^x\\ [2-e^x]f(x)&=4-e^x\\ f(x)&=\frac{4-e^x}{2-e^x} \end{align*}2.4.7.35. (✳).
Solution.
for some constant \(d\text{.}\) At time \(t=0\text{,}\) the height is \(y=2\text{,}\) so \(d=\displaystyle\frac{\pi}{c}\cdot\frac{2^{2p+{1\over 2}}}{2p+{1\over 2}}\text{.}\)
\begin{align*} t&=\frac{\pi}{c}\bigg(\frac{2^{2p+{1\over 2}}}{2p+{1\over 2}} -\frac{y^{2p+{1\over 2}}}{2p+{1\over 2}}\bigg)\\ &=\frac{\pi}{c(2p+\frac12)}\left(2^{2p+\frac12}-y^{2p+\frac12}\right) \end{align*}2.4.7.36.
Solution.
- If we let \(f(t)=0\) for all \(t\text{,}\) then its average over any interval is 0, as is its root mean square.
- Let’s start by simplifying the given equation.\begin{align*} \frac{1}{x-a}\int_a^x f(t)\,\dee{t}&=\sqrt{\frac{1}{x-a}\int_a^x f^2(t)\,\dee{t}}\\ \frac{1}{\sqrt{x-a}}\int_a^x f(t)\,\dee{t}&=\sqrt{\int_a^x f^2(t)\,\dee{t}} \tag{E1}\\ \color{red}{\diff{}{x}\left\{\frac{1}{\sqrt{x-a}}\int_a^x f(t)\,\dee{t}\right\}} &=\color{blue}{\diff{}{x}\left\{\sqrt{\int_a^x f^2(t)\,\dee{t}}\right\}} \tag{E2} \end{align*}For the derivative on the left, we use the product rule and the Fundamental Theorem of Calculus, part 1.\begin{align*} \amp\color{red}{\diff{}{x}\left\{\frac{1}{\sqrt{x-a}}\int_a^x f(t)\,\dee{t}\right\}}\\ \amp\hskip0.1in=\diff{}{x}\left\{\frac{1}{\sqrt{x-a}}\right\}\int_a^x f(t)\,\dee{t} + \frac{1}{\sqrt{x-a}}\cdot\diff{}{x}\left\{\int_a^x f(t)\,\dee{t}\right\}\\ &\hskip0.1in=-\frac{1}{2\sqrt{x-a}^3}\int_a^x f(t)\,\dee{t} + \frac{f(x)}{\sqrt{x-a}}\\ &\hskip0.1in=\color{red}{ \frac{1}{\sqrt{x-a}}\left[f(x) - \frac{1}{2(x-a)}\int_a^x f(t)\,\dee{t}\right]} \end{align*}For the derivative on the right in Equation (E2) we use the chain rule and the Fundamental Theorem of Calculus part 1\begin{align*} \color{blue}{ \diff{}{x}\left\{\sqrt{\int_a^x f^2(t)\,\dee{t}}\right\} }&=\frac{1}{2}\left(\int_a^x f^2(t)\,\dee{t}\right)^{-\frac12}\!\!\cdot\diff{}{x}\left\{\int_a^x f^2(t)\,\dee{t}\right\}\\ &=\color{blue}{\frac{f^2(x)}{2\sqrt{\int_a^xf^2(t)\,\dee{t}}}} \end{align*}So, Equation (E2) yields the following:\begin{equation*} \color{red}{ \!\!\frac{1}{\sqrt{x\!-\!a}}\left[f(x) - \frac{1}{2(x\!!a)}\int_a^x\!\! f(t)\,\dee{t}\right]} =\color{blue}{ \frac{f^2(x)}{2\sqrt{\int_a^xf^2(t)\,\dee{t}}}} \tag{E3} \end{equation*}
- From Equation (E1), \(\sqrt{\int_a^x f^2(t)\,\dee{t}} = \frac{1}{\sqrt{x-a}}\int_a^x f(t)\,\dee{t}\text{.}\)\begin{equation*} \frac{1}{\sqrt{x-a}}\left[f(x) - \frac{1}{2(x-a)}\int_a^x f(t)\,\dee{t}\right]=\frac{f^2(x)}{2\frac{1}{\sqrt{x-a}}\int_a^x f(t)\,\dee{t}} \end{equation*}and\begin{equation*} \frac{2}{x-a}\int_a^x f(t)\,\dee{t}\left[f(x) - \frac{1}{2(x-a)}\int_a^x f(t)\,\dee{t}\right]=f^2(x) \end{equation*}
-
Now what we have is a differential equation, although it might not look like it. Let \(Y(x) = \int_a^x f(t)\,\dee{t}\text{.}\) Then \(\diff{Y}{x}(x) = f(x)\text{.}\)\begin{align*} \frac{2}{x-a}Y\left[\diff{Y}{x} - \frac{1}{2(x-a)}Y\right]&=\left(\diff{Y}{x}\right)^2 \tag{E4} \end{align*}We’re used to solving differential equations of the form \(\diff{Y}{x}=\)(something). So, let’s manipulate our equation until it has this form.\begin{gather*} \left(\diff{Y}{x}\right)^2-\left(\frac{2Y}{x-a}\right)\left(\diff{Y}{x}\right)+\left(\frac{Y}{x-a}\right)^2=0 \end{gather*}This is a quadratic equation, with variable \(\diff{Y}{x}\text{.}\) Its solutions are:\begin{align*} \diff{Y}{x}&=\frac{\left(\frac{2Y}{x-a}\right)\pm\sqrt{\left(\frac{2Y}{x-a}\right)^2-4\cdot\left(\frac{Y}{x-a}\right)^2}}{2}\\ &=\frac{\frac{2Y}{x-a}\pm 0}{2}\\ &=\frac{Y}{x-a} \end{align*}This gives us the separable differential equation\begin{align*} \diff{Y}{x}&=\frac{Y}{x-a}\\ \frac{\dee{Y}}{Y}&=\frac{\dee{x}}{x-a} \tag{E5}\\ \int\frac{\dee{Y}}{Y}&=\int\frac{\dee{x}}{x-a}\\ \log|Y|&=\log|x-a|+C\\ |Y|&=e^{\log|x-a|+C} = |x-a|e^C\\ Y&=D(x-a) \end{align*}where \(D\) is some constant, \(e^C\) or \(-e^C\text{.}\) Note this covers all real constants except \(D=0\text{.}\) If \(D=0\text{,}\) then \(Y(x)=0\) for all \(x\text{.}\) This function also satisfies Equation (E4), so indeed,\begin{equation*} Y(x)=D(x-a) \tag{E6} \end{equation*}for any constant \(D\) is the family of equations satisfying our differential equation.Remark: the reason we “lost” the solution \(Y(x)=0\) is that in Equation (E5), we divided by \(Y\text{,}\) thus tacitly assuming it was not identically 0.
-
Remember \(Y=\int_a^x f(t)\,\dee{t}\text{.}\) So, Equation (E6) tells us:\begin{align*} \int_a^x f(t)\,\dee{t}&=D(x-a)\\ \diff{}{x}\left\{\int_a^x f(t)\,\dee{t}\right\}&=\diff{}{x}\{D(x-a)\}\\ f(x)&=D \end{align*}We should check that this function works.\begin{align*} f_{\text{avg}} &= \frac{1}{x-a}\int_a^x D\,\dee{t} = \frac{1}{x-a}\Big[Dt\Big]_{t=a}^{t=x} = \frac{Dx-Da}{x-a}=D\\ f_{\text{RMS}} &= \sqrt{\frac{1}{x-a}\int_a^x D^2\,\dee{t}} =\sqrt{\frac{1}{x-a}\Big[D^2t\Big]_{t=a}^{t=x}}\\ \amp=\sqrt{\frac{D^2x-D^2a}{x-a}}=\sqrt{D^2}=|D| \end{align*}So, \(f(x)=D\) works only if \(D\) is nonnegative.That is: the only functions whose average matches their root mean square over every interval are constant, nonnegative functions.Remark: it was step (c) where we introduced the erroneous answer \(f(x)=D\text{,}\) \(D \lt 0\) to our solution. In Equation (E3), \(f(x)=D\) is not a solution if \(D \lt 0\text{:}\)\begin{align*} \frac{1}{\sqrt{x-a}}\left[f(x) - \frac{1}{2(x-a)}\int_a^x f(t)\,\dee{t}\right]&=\frac{f^2(x)}{2\sqrt{\int_a^x f^2(t)\,\dee{t}}}\\ \frac{1}{\sqrt{x-a}}\left[D - \frac{1}{2(x-a)}\int_a^x D\,\dee{t}\right]&=\frac{D^2}{2\sqrt{\int_a^x D^2\,\dee{t}}}\\ \frac{1}{\sqrt{x-a}}\left[D - \frac{1}{2(x-a)}D(x-a)\right]&=\frac{D^2}{2\sqrt{ D^2(x-a)}}\\ \frac{1}{\sqrt{x-a}}\left[\frac{1}{2}D \right]&=\frac{D^2}{2|D|\sqrt{x-a}}\\ D&=\frac{D^2}{|D|}=|D| \end{align*}In (c), we replace \(\sqrt{\int_a^x f^2(t)\,\dee{t}}\text{,}\) which cannot be negative, with \(\frac{1}{\sqrt{x-a}}\int_a^x f(t)\,\dee{t}\text{,}\) which could be negative if \(f(t)=D \lt 0\text{.}\) Indeed, if \(f(t)=D\text{,}\) then \(\sqrt{\int_a^x f^2(t)\,\dee{t}} = |D|\sqrt{x-a}\text{,}\) while \(\frac{1}{\sqrt{x-a}}\int_a^x f(t)\,\dee{t}=D\sqrt{x-a}\text{.}\) It is at this point that negative functions creep into our solution.
2.4.7.37.
Solution.
The right integral is in exactly the form we would use for a change of variables (substitution) to \(y\text{.}\)
\begin{align*} \diff{y}{x}&=\int\left(\frac{2}{y^3}\right)\,\dee{y} = -\frac{1}{y^2}+C\\ \end{align*}When \(y=1\text{,}\) \(\diff{y}{x}=3\text{.}\)
\begin{align*} 3& =-\frac{1}{1}+C\\ C&=4\\ \end{align*}So,
\begin{align*} \diff{y}{x}&=-\frac{1}{y^2}+4\\ \end{align*}This is a separable differential equation.
\begin{align*} \diff{y}{x}&=\frac{4y^2-1}{y^2}\\ \frac{y^2}{4y^2-1}\,\dee{y}&=\dee{x}\\ \int \frac{y^2}{4y^2-1}\,\dee{y}&=\int \dee{x}\tag{$*$} \end{align*}When \(x=-\frac{1}{16}\log 3\text{,}\) \(y=1\text{.}\)
\begin{align*} -\frac{1}{16}\log 3 +C &=\frac{1}{4}\left(1+\frac14\log\left| \frac{2-1}{2+1}\right|\right) = \frac{1}{4}+\frac{1}{16}\log\frac{1}{3}\\ C&=\frac{1}{4}\\ \end{align*}So,
\begin{align*} x+\frac{1}{4}&=\frac{1}{4}\left(y+\frac{1}{4}\log\left|\frac{2y-1}{2y+1}\right|\right)\\ x&=\frac{1}{4}\left(y-1+\frac{1}{4}\log\left|\frac{2y-1}{2y+1}\right|\right) \end{align*}Differentiating with respect to \(x\) again, using the chain rule,
\begin{align*} \ddiff{2}{y}{x}&=\frac{2}{y^3}\cdot\diff{y}{x} \end{align*}
3 Sequences and series
3.1 Sequences
3.1.2 Exercises
3.1.2.1.
Solution.
3.1.2.2.
Solution.
3.1.2.3.
Solution.
3.1.2.4.
Solution.
3.1.2.5.
Solution.
3.1.2.6.
Solution.
3.1.2.7.
Solution.
- Since \(-1 \leq \sin n \leq 1\) for all \(n\text{,}\) one potential set of upper and lower bound is\begin{equation*} \dfrac{-1}{n}\le \dfrac{\sin n}{n}\le \dfrac{1}{n} \end{equation*}Note \(\lim\limits_{n \to \infty}\dfrac{-1}{n} = \lim\limits_{n \to \infty}\dfrac{1}{n}\text{,}\) so these are valid comparison sequences for the squeeze theorem.
- Since \(\textcolor{red}{-1\leq \sin n \leq 1}\) and \(\textcolor{blue}{-5 \leq -5\cos n \leq 5}\) for all \(n\text{,}\) we see\begin{align*} 7\textcolor{red}{-1}\textcolor{blue}{-5}&\leq 7+\textcolor{red}{\sin n} -\textcolor{blue}{ 5 \cos n} \leq 7+\textcolor{red}1+\textcolor{blue}5\\ 1&\leq 7+\textcolor{red}{\sin n} -\textcolor{blue}{ 5 \cos n} \leq13 \end{align*}This gives us the idea to try the bounds\begin{equation*} \dfrac{n^2}{13e^n} \le \dfrac{n^2}{e^n(7+\sin n - 5\cos n)} \le \dfrac{n^2}{e^n} \end{equation*}We check that \(\lim\limits_{n \to \infty}\dfrac{n^2}{13e^n} = \lim\limits_{n \to \infty}\dfrac{n^2}{e^n} \) (they’re both 0 — you can verify using l’Hôpital’s rule), so these are indeed reasonable bounds to choose to use with the squeeze theorem. Alternatively, since \(0\le \frac{n^2}{13e^n}\text{,}\) we can also use\begin{equation*} 0 \le \dfrac{n^2}{e^n(7+\sin n - 5\cos n)} \le \dfrac{n^2}{e^n} \end{equation*}
- Since \((-n)^{-n} = \dfrac{1}{(-n)^n} = \dfrac{(-1)^n}{n^n}\text{,}\) we see\begin{equation*} \dfrac{-1}{n^n} \le (-n)^{-n} \le \dfrac{1}{n^n} \end{equation*}Since both \(\lim\limits_{n \to \infty}\dfrac{-1}{n^n} \) and \(\lim\limits_{n \to \infty}\dfrac{1}{n^n} \) are 0, these are reasonable bounds to use with the squeeze theorem.
3.1.2.8.
Solution.
- Note \(a_n=b_n\) since \(n=|n|\) for all \(n\ge 1\text{.}\) Then \(a_n=b_n=1+\dfrac{1}{n}=\dfrac{n+1}{n}\text{.}\) So, whenever \(n\) is a whole number, \(a_n\) and \(b_n\) are the same as \(h(n)\) and \(i(n)\text{.}\) (Be careful here: \(h(x) \neq i(x)\) when \(x\) is not a whole number.)
- \(\displaystyle c_n=e^{-n}=\dfrac{1}{e^n}=j(n)\)
- For any integer \(n\text{,}\) \(\cos(\pi n) = (-1)^n\text{.}\) So, \(d_n = f(n)\text{.}\)
- Similarly, \(e_n=g(n)\text{.}\)
-
According to Theorem 3.1.6, if any of the functions on the right have limits that exist as \(x \to \infty\text{,}\) then these limits match the limits of their corresponding sequences. So, we only have to be suspicious of \(f(x)\) and \(i(x)\text{,}\) since these do not converge.The limit \(\lim\limits_{x \to \infty}f(x)\) does not exist, and \(f(n)=d_n\text{;}\) the limit \(\lim\limits_{n \to \infty}d_n\) also does not exist. (We generally don’t write equality for two things that don’t exist: equality refers to numerical value, and these have none.)
4
The idea “two things that both don’t exist are equal” is also rejected because it can lead to contradictions. For example, in the real numbers \(\sqrt{-1} \) and \(\sqrt{-2}\) don’t exist; if we write \(\sqrt{-1}=\sqrt{-2}\text{,}\) then squaring both sides yields the inanity \(-1=-2\text{.}\)The limit \(\lim\limits_{x \to \infty}i(x)\) does not exist, because \(i(x)=0\) when \(x\) is not a whole number, while \(i(x)\) approaches 1 when \(x\) is a whole number. However, \(\lim\lim\limits_{n \to \infty}a_n=\lim\limits_{n \to \infty}b_n=1\text{.}\)So, using our answers from part (a), we match the following:- \(\displaystyle \lim\limits_{n \to \infty}a_n=\lim\limits_{n \to \infty}b_n=\lim\limits_{x \to \infty}h(x)=1\)
- \(\displaystyle \lim\limits_{n \to \infty}c_n=\lim\limits_{n \to \infty}e_n=\lim\limits_{x \to \infty}g(x)=\lim\limits_{x \to \infty}j(x)=0\)
- \(\lim\limits_{n \to \infty} d_n\text{,}\) \(\lim\limits_{x\rightarrow\infty} f(x)\) and \(\lim\limits_{x\rightarrow\infty} i(x)\) do not exist.
3.1.2.9.
Solution.
-
Solution 1: One way to do that is to remember that \(\pi\) is somewhat close to \(\dfrac{22}{7}\text{.}\) Then when we multiply \(\pi\) by a multiple of 7, we should get something close to an integer. In particular, \(7\pi\text{,}\) \(21\pi\text{,}\) and \(35\pi\) should be reasonably close to \(7\left(\dfrac{22}{7}\right)=22\text{,}\) \(21\left(\dfrac{22}{7}\right)=66\text{,}\) and \(35\left(\dfrac{22}{7}\right)=110\text{,}\) respectively. We check whether they are close enough:\begin{align*} 7\pi&\approx 21.99 &\qquad 21\pi &\approx 65.97 & \qquad 35\pi &\approx 109.96 \end{align*}So indeed, \(22\text{,}\) \(66\text{,}\) and \(110\) are all within 0.1 of some odd multiple of \(\pi\text{.}\)Since the cosine of an odd multiple of \(\pi\) is \(-1\text{,}\) we expect all of the sequence values to be close to \(-1\text{.}\) Using a calculator:\begin{align*} a_{22} &= \cos(22) \approx -0.99996,\\ a_{66} &= \cos(66) \approx -0.99965,\\ a_{110} &= \cos(110) \approx -0.99902 \end{align*}
-
Solution 2: Alternately, we could have just listed odd multiple of \(\pi\) until we found three that are close to integers.\begin{equation*} \begin{array}{c|c} \hline \mathbf{2k+1}&\mathbf{(2k+1)}\pmb{\pi}\\ \hline 1&3.14\\ 3&9.42\\ 5&15.71\\ 7&\textcolor{red}{21.99}\\ 9&28.27\\ 11&34.56\\ 13&40.84\\ 15&47.12\\ 17&53.41\\ 19&59.69\\ 21&\textcolor{red}{65.97}\\ 23&72.26\\ 25&78.54\\ 27&84.82\\ 29&91.11\\ 31&97.39\\ 33&103.67\\ 35&\textcolor{red}{109.96} \end{array} \end{equation*}Some earlier odd multiples of \(\pi\) (like \(15\pi\) and \(29\pi\)) get fairly close to integers, but not within 0.1.
-
Solution 1:\begin{align*} \dfrac{2k+1}{2}\pi&\approx \frac{(2k+1)\times 22}{2\times 7}=11\times\frac{2k+1}{7} \end{align*}So, we expect our values to be close to integers when \(2k+1\) is a multiple of 7. For example, \(2k+1=7\text{,}\) \(2k+1=21\text{,}\) and \(2k+1=35\text{.}\)We check:\begin{equation*} \begin{array}{l| l| l} \mathbf{x}&\mathbf{n} & \mathbf{a_n}\\ \hline 7\times \dfrac{\pi\vphantom{^A}}{2}\approx 10.99557 & 11&a_{11}\approx0.0044 \\ 21\times \dfrac{\pi}{2} \approx 32.98672&33&a_{33}\approx-0.0133 \\ 35\times \dfrac{\pi}{2} \approx 54.97787&55& a_{55}\approx 0.0221 \\ \end{array} \end{equation*}These seem like values of \(a_n\) that are all pretty close to 0.
-
Solution 2: We could have listed the first several values of \(a_n\text{,}\) and looked for some that are close to 0.\begin{equation*} \begin{array}{c|c} \mathbf{n}&\mathbf{a_n}\\ \hline 1&0.54\\ 2&-0.42\\ 3&-0.99\\ 4&-0.65\\ 5&0.28\\ 6&0.96\\ 7&0.75\\ 8&-0.15\\ 9&-0.91\\ 10&-0.84 \end{array} \end{equation*}Oof. Nothing very close yet. Maybe a better way is to list values of \(\frac{2k+1}{2}\pi\text{,}\) and see which ones are close to integers.\begin{equation*} \begin{array}{c|c} \mathbf{2k+1}&\mathbf{\frac{2k+1}{2}}\pmb{\pi}\\ \hline 1&1.57\\ 3&4.71\\ 5&7.85\\ 7&\textcolor{red}{10.996}\\ 9&14.14\\ 11&17.28\\ 13&20.42\\ 15&23.56\\ 17&26.70\\ 19&29.85\\ 21&\textcolor{red}{32.99}\\ 23&36.13\\ 25&39.27\\ 27&42.41\\ 29&45.55\\ 31&48.69\\ 33&51.84\\ 35&\textcolor{red}{54.98} \end{array} \end{equation*}We find roughly the same candidates we did in Solution 1, depending on what we’re ready to accept as “close”.
- Let, for each integer \(k\geq 1\text{,}\) \(n_k\) be the integer that is closest to \(2k\pi\text{.}\) Then \(2k\pi-\frac{1}{2}\leq n_k \leq 2k\pi+\frac{1}{2}\) so that \(\cos(n_k)\geq\cos\frac{1}{2}\geq 0.8\text{.}\) Consequently, if \(\lim\limits_{n \to \infty}\cos n=c\) exists, we must have \(c\geq 0.8\text{.}\)
- Let, for each integer \(k\geq 1\text{,}\) \(n'_k\) be the integer that is closest to \((2k+1)\pi\text{.}\) Then \((2k+1)\pi-\frac{1}{2}\leq n'_k \le (2k+1)\pi+\frac{1}{2}\) so that \(\cos(n'_k)\leq-\cos\frac{1}{2}\leq -0.8\text{.}\) Consequently, if \(\lim\limits_{n \to \infty}\cos n=c\) exists, we must have \(c\leq -0.8\text{.}\)
- It is impossible to have both \(c\geq 0.8\) and \(c\leq -0.8\text{,}\) so \(\lim\limits_{n \to \infty}\cos n\) does not exist.
3.1.2.10.
Solution.
-
Since the numerator has a higher degree than the denominator, this sequence will diverge to positive or negative infinity; since its terms are positive for large \(n\text{,}\) its limit is (positive) infinity. (You can imagine that the numerator is growing much, much faster than the denominator, leading the terms to have a very, very large absolute value.)Calculating the longer way:\begin{align*} a_n &= \dfrac{3n^2-2n+5}{4n+3}\left(\dfrac{\frac1n}{\frac1n}\right)=\dfrac{3n-2+\frac5n}{4+\frac3n}\\ \lim_{n \to \infty}a_n &= \lim_{n \to \infty}\dfrac{3n-2+\frac5n}{4+\frac3n} = \lim_{n \to \infty}\dfrac{3n-2+0}{4+0} = \infty \end{align*}
-
Since the numerator has the same degree as the denominator, as \(n\) goes to infinity, this sequence will converge to the ratio of their leading coefficients: \(\dfrac{3}{4}\text{.}\) (You can imagine that the numerator is growing at roughly the same rate as the denominator, so the terms settle into an almost-constant ratio.)Calculating the longer way:\begin{align*} b_n &= \dfrac{3n^2-2n+5}{4n^2+3}\left(\dfrac{\frac{1}{n^2}}{\frac{1}{n^2}}\right)=\dfrac{3-\frac2n+\frac{5}{n^2}}{4+\frac{3}{n^2}}\\ \lim_{n \to \infty}b_n &= \lim_{n \to \infty}\dfrac{3-\frac2n+\frac{5}{n^2}}{4+\frac{3}{n^2}} = \frac{3-0+0}{4+0}=\frac{3}{4} \end{align*}
-
Since the numerator has a lower degree than the denominator, this sequence will converge to 0 as \(n\) goes to infinity. (You can imagine that the denominator is growing much, much faster than the numerator, leading the terms to be very, very small.)Calculating the longer way:\begin{align*} c_n &= \dfrac{3n^2-2n+5}{4n^3+3}\left(\dfrac{\frac{1}{n^3}}{\frac{1}{n^3}}\right)=\dfrac{\frac{3}{n}-\frac{2}{n^2}+\frac{5}{n^3}}{4+\frac{3}{n^3}}\\ \lim_{n \to \infty}c_n &= \lim_{n \to \infty}\dfrac{\frac{3}{n}-\frac{2}{n^2}+\frac{5}{n^3}}{4+\frac{3}{n^3}} = \dfrac{0-0+0}{4+0}=0 \end{align*}
3.1.2.11.
Solution.
Since \(e \lt 3\text{,}\) we see \(3-e\) is positive, so \(\lim\limits_{n \to \infty}n^{3-e}=\infty\text{.}\)
\begin{align*} \lim_{n \to \infty}a_n&=\lim_{n \to \infty} \dfrac{4n^{3-e}-\frac{21}{n^e}}{1+\frac{1}{n^{e+1}}} =\lim_{n \to \infty} \dfrac{4n^{3-e}-0}{1+0} = \infty \end{align*}3.1.2.12.
Solution.
3.1.2.13.
Solution.
The denominator grows without bound, so \(\lim\limits_{n \to \infty}\dfrac{\cos(n+n^2)}{n}=0\text{.}\)
The denominator grows without bound, and the numerator never gets very large, so \(\lim\limits_{n \to \infty}\dfrac{\cos(n+n^2)}{n}=0\text{.}\)
- Since \(-1\leq \cos(n+n^2) \leq 1\) for all \(n\text{,}\) we choose functions \(a_n = \frac{-1}{n}\) and \(b_n = \frac{1}{n}\text{.}\) Then \(a_n \leq c_n \leq b_n\) for all \(n\text{.}\)
- Both \(\lim\limits_{n \to \infty}a_n=0\) and \(\lim\limits_{n \to \infty}b_n=0\text{.}\)
3.1.2.14.
Solution.
- Since \(-1 \leq \sin n \leq 1\) for all \(n\text{,}\) let \(b_n = \frac{n^{-1}}{n^2} = \frac{1}{n^3}\) and \(c_n = \frac{n}{n^2}=\frac{1}{n}\text{.}\) Then \(b_n \leq a_n \leq c_n\text{.}\)
- Both \(\lim\limits_{n \to \infty}b_n=0\) and \(\lim\limits_{n \to \infty}c_n=0\text{.}\)
3.1.2.15.
Solution.
3.1.2.16.
Solution.
- Solution 1: Let’s use the squeeze theorem. Since \(\textcolor{blue}{\sin (n^2)}\) and \(\textcolor{red}{\sin n}\) are both between \(-1\) and 1 for all \(n\text{,}\) we note:\begin{align*} 1 +3\textcolor{blue}{(-1)}-2\textcolor{red}{(1)}&\leq 1+3\,\textcolor{blue}{\sin(n^2)}-2\,\textcolor{red}{\sin n} \leq 1 +3\textcolor{blue}{(1)}-2\textcolor{red}{(-1)}\\ -4&\leq 1+3\,{\sin(n^2)}-2\,{\sin n} \leq 6 \end{align*}This allows us to choose suitable bounding functions for the squeeze theorem.
- Let \(b_n = -\dfrac{4}{n}\) and \(c_n = \dfrac{6}{n}\text{.}\) From the work above, we see \(b_n \leq a_n \leq c_n\) for all \(n\text{.}\)
- Both \(\lim\limits_{n \to \infty}b_n=0\) and \(\lim\limits_{n \to \infty}c_n=0\text{.}\)
So, by the squeeze theorem, \(\lim\limits_{n \to \infty} \dfrac{1+3\sin(n^2)-2\sin n}{n}=0\text{.}\) -
Solution 2: We simplify slightly to begin.\begin{align*} a_n &= \dfrac{1+3\sin(n^2)-2\sin n}{n} = \dfrac{1}{n}+3\cdot\frac{\sin (n^2)}{n} - 2\cdot\frac{\sin n}{n} \end{align*}We apply the squeeze theorem to the pieces \(\textcolor{blue}{\dfrac{\sin (n^2)}{n} }\) and \(\textcolor{red}{\dfrac{\sin n}{n}}\text{.}\)
- Let \(b_n = \dfrac{-1}{n}\) and \(c_n = \dfrac{1}{n}\text{.}\) Then \(\textcolor{blue}{b_n \leq \dfrac{\sin (n^2)}{n} \leq c_n}\text{,}\) and \(\textcolor{red}{ b_n \leq \dfrac{\sin n}{n} \leq c_n }\text{.}\)
- Both \(\lim\limits_{n \to \infty} b_n=0\) and \(\lim\limits_{n \to \infty} c_n=0\text{.}\)
So, by the squeeze theorem, \(\textcolor{blue}{ \lim\limits_{n \to \infty}\dfrac{\sin (n^2)}{n} =0 }\) and \(\textcolor{red}{ \lim\limits_{n \to \infty}\dfrac{\sin n}{n}=0 }\text{.}\)Now, using the arithmetic of limits from Theorem 3.1.8,\begin{align*} \lim_{n \to \infty}a_n &=\lim_{n \to \infty}\left[ \dfrac{1}{n}+3\cdot\frac{\sin (n^2)}{n} - 2\cdot\frac{\sin n}{n}\right]\\ &=0+3\cdot 0 - 2\cdot 0 =0 \end{align*}
3.1.2.17.
Solution.
-
Solution 1: Let’s try dividing the numerator and denominator by \(2^n\) (the dominant term in the denominator; this is the same idea behind factoring out the leading term in rational expressions).\begin{equation*} b_n = \frac{e^n}{2^n+n^2}\left(\frac{\frac{1}{2^n}}{\frac{1}{2^n}}\right) = \frac{\left(\frac{e}{2}\right)^n}{1+\frac{n^2}{2^n}} \end{equation*}Since \(e \gt 2\text{,}\) we see \(\dfrac{e}{2} \gt 1\text{,}\) and so \(\lim\limits_{n \to \infty}\left(\dfrac{e}{2}\right)^n=\infty\text{.}\) Since exponential functions grow much, much faster than polynomial functions, we also see \(\lim\limits_{n \to \infty}\frac{n^2}{2^n}=0\text{.}\) So,\begin{equation*} \lim\limits_{n \to \infty}b_n = \lim\limits_{n \to \infty} \frac{\left(\frac{e}{2}\right)^n}{1+\frac{n^2}{2^n}} = \lim\limits_{n \to \infty} \frac{\left(\frac{e}{2}\right)^n}{1+0} = \infty \end{equation*}
- Solution 2: Since the numerator and denominator both increase without bound, we apply l’Hôpital’s rule. Recall \(\diff{}{x}\{2^x\} = 2^x\log 2\text{.}\)\begin{align*} \lim_{n \to \infty}b_n&=\lim_{n \to \infty}\underbrace{\dfrac{e^n}{2^n+n^2}}_{\atp{\mathrm{num}\to \infty}{\mathrm{den}\to\infty}}\\ &=\lim_{n \to \infty}\underbrace{\frac{e^n}{2^n\log 2 + 2n}}_{\atp{\mathrm{num}\to \infty}{\mathrm{den}\to\infty}}\\ &=\lim_{n \to \infty} \underbrace{\frac{e^n}{2^n(\log 2)^2 + 2}}_{\atp{\mathrm{num}\to\infty}{\mathrm{den}\to\infty}}\\ &=\lim_{n \to \infty}\frac{e^n}{2^n(\log 2)^3}\\ &=\frac{1}{(\log 2)^3}\lim_{n \to \infty}{\left(\frac{e}{2}\right)^n}\\ &=\infty \end{align*}Since \(e \gt 2\text{,}\) we see \(\dfrac{e}{2} \gt 1\text{,}\) and so \(\lim\limits_{n \to \infty}\left(\dfrac{e}{2}\right)^n=\infty\text{.}\)
3.1.2.18. (✳).
Solution.
- \(-1 \leq \sin k \leq 1\) for all \(k\text{,}\) so \(-1 \leq \sin^3k \leq 1\text{.}\) Let \(b_k = \frac{-1}{k+1}\) and \(c_k = \frac{1}{k+1}\text{.}\) Then \(b_k \leq a_k \leq c_k\text{.}\)
- Both \(\lim\limits_{k \to \infty}b_k=0\) and \(\lim\limits_{k \to \infty}c_k=0\text{.}\)
3.1.2.19. (✳).
Solution.
- Let \(a_n=-\sin\left(\frac{1}{n}\right)\) and \(b_n=\sin\left(\frac{1}{n}\right)\text{.}\) Then \(a_n\leq (-1)^n\sin\left(\frac{1}{n}\right) \leq b_n\text{.}\)
- Both \(\lim\limits_{n \to \infty} -\sin\left(\frac{1}{n}\right)=0\) and \(\lim\limits_{n \to \infty} \sin\left(\frac{1}{n}\right)=0\text{,}\) since \(\lim\limits_{n \to \infty} \frac{1}{n}=0\) and \(\sin 0=0.\)
3.1.2.20. (✳).
Solution.
3.1.2.21. (✳).
Solution.
If \(n \to \infty\text{,}\) then \(x = \frac{1}{n}\to 0\text{.}\)
\begin{align*} \lim_{n \to \infty}2n\sin\left(\frac{1}{n}\right)&=2\lim_{x \to 0}\frac{\sin x}{x}\\ \end{align*}That limit is familiar:
\begin{align*} &=2(1)=2\\ \end{align*}Then:
\begin{align*} \lim_{n \to \infty}\log\left(2n\sin\left(\frac{1}{n}\right)\right)&=\log 2 \end{align*}3.1.2.22.
Solution.
Now, we’ll cancel out \(n\) from the top and the bottom. Note \(n=\sqrt{n^2}\text{.}\)
\begin{align*} &=\frac{10n}{\sqrt{n^2+5n}+\sqrt{n^2-5n}}\left(\frac{\frac1n}{\frac1n}\right)\\ &=\frac{10n}{\sqrt{n^2+5n}+\sqrt{n^2-5n}}\left(\frac{\frac1n}{\frac{1}{\sqrt{n^2}}}\right)\\ &=\frac{10}{\sqrt{1+\frac5n}+\sqrt{1-\frac5n}}\\ \end{align*}Now, the limit is clear.
\begin{align*} \lim_{n \to \infty}\frac{10}{\sqrt{1+\frac5n}+\sqrt{1-\frac5n}}&=\frac{10}{\sqrt{1+0}+\sqrt{1+0}}=\frac{10}{1+1}=5 \end{align*}3.1.2.23.
Solution.
Now, the limit is clear.
\begin{align*} \lim_{n \to \infty}\left[\sqrt{n^2+5n}-\sqrt{2n^2-5}\right]&= \lim_{n \to \infty}\left[n\left(\sqrt{1+\frac{5}{n}} - \sqrt{2-\frac{5}{n^2}}\right) \right]\\ &=\lim_{n \to \infty}\left[n\left(\sqrt{1+0} - \sqrt{2-0}\right) \right]\\ &=\lim_{n \to \infty}\left[n\left(-1\right) \right]=-\infty \end{align*}3.1.2.24.
Solution.
This reminds us of the definition of a derivative.
\begin{align*} \diff{}{x}\left\{x^{100}\right\}&=\lim_{h \to 0}\frac{(x+h)^{100}-x^{100}}{h} \end{align*}3.1.2.25.
Solution.
We want \(n\to \infty\text{,}\) so we set \(h = \frac{1}{n}\text{.}\)
\begin{align*} &=\lim_{\frac{1}{n}\to 0}\frac{f\left(a+\frac{1}{n}\right)-f(a)}{\frac{1}{n}}\\ &=\lim_{n\to \infty}n\left[f\left(a+\frac{1}{n}\right)-f(a)\right]\\ \end{align*}We also could have chosen \(h=-\frac{1}{n}\text{,}\) which leads to the following:
\begin{align*} \lim_{h \to 0}\frac{f(a+h)-f(a)}{h}&= \lim_{-\frac1n \to 0}\frac{f\left(a-\frac{1}{n}\right)-f(a)}{-1/n}\\ &=\lim_{n \to \infty}-n\left(f\left(a-\frac{1}{n}\right)-f(a)\right)\\ &=\lim_{n \to \infty}n\left(f(a)-f\left(a-\frac{1}{n}\right)\right) \end{align*}3.1.2.26.
Solution.
3.1.2.27.
Solution.
-
\(\displaystyle f_2(x) = \begin{cases} 1 & 2 \leq x \lt 3\\ 0 & \text{else} \end{cases}\)
-
\(\displaystyle f_3(x) = \begin{cases} 1 & 3 \leq x \lt 4\\ 0 & \text{else} \end{cases}\)
-
For any \(n\text{,}\) \(f_n(x)=1\) for an interval of length 1, and \(f_n(x)=0\) for all other \(x\text{.}\) So, the area under the curve is a square of side length one.Then \(A_n = \int_0^\infty f_n(x)\,\dee{x}=1\) for all \(n\text{.}\) That is, the sequence \(\{A_n\}\) is simply \(\{1,1,\ldots,1\}\text{,}\) a sequence of all 1s.
- Given the description above, \(\displaystyle\lim_{n \to \infty}A_n=1\text{.}\)
- For any fixed \(x\text{,}\) recall \(\{f_n(x)\} = \{0,\ldots,0,1,0,\ldots 0,0,0,0,0,\ldots\}\text{.}\) In particular, there are infinitely many zeroes at its end. So, \(\displaystyle\lim_{n \to \infty} f_n(x)=0\text{.}\) Then \(g(x)=0\) for every \(x\text{.}\)
- Given the description above, \(\displaystyle \int_0^\infty g(x)\,\dee{x} = \int_0^{\infty}0\,\dee{x}=0\text{.}\)
3.1.2.28.
Solution.
- Solution 1: Define \(x=\frac{1}{n}\text{,}\) and \(f(x)=\left(1+3x+5x^2\right)^{1/x}\text{.}\) Then \(b_n=f\left(\frac{1}{n}\right)=f(x)\text{,}\) and\begin{equation*} \lim_{n \to \infty} f\left(\frac{1}{n}\right) = \lim_{x \to 0^+}f(x). \end{equation*}If this limit exists, it is equal to \(\lim\limits_{n \to \infty}b_n\text{.}\)\begin{align*} \lim_{x \to 0^+}f(x)&= \lim_{x \to 0^+} \left(1+3x+5x^2\right)^{1/x}\\ \lim_{x \to 0^+}\log[f(x)]&= \lim_{x \to 0^+} \log\left[\left(1+3x+5x^2\right)^{1/x}\right]\\ \amp=\underbrace{ \lim_{x \to 0^+}\frac{ \log\left[1+3x+5x^2 \right]}{x}}_{\atp{\mathrm{num}\to 0}{\mathrm{den}\to0}}\\ &=\lim_{x \to 0^+}\frac{\frac{3+10x}{1+3x+5x^2}}{1}=3\\ \lim_{x \to 0^+}f(x)&=e^3 \end{align*}Since the limit exists, \(\lim\limits_{n \to \infty}b_n=e^3\text{.}\)
-
Solution 2: If we didn’t see the nice simplifying trick of letting \(x=\frac{1}{n}\text{,}\) we can still solve the problem using \(g(x)=\left(1+\frac{3}{x}+\frac{5}{x^2}\right)^{x}\text{:}\)\begin{align*} g(x)&=\left(1+\frac{3}{x}+\frac{5}{x^2}\right)^x\\ \log [g(x)]&= x\log\left[1+\frac{3}{x}+\frac{5}{x^2}\right] =\underbrace{\frac{\log\left[1+\frac{3}{x}+\frac{5}{x^2}\right]}{1/x}}_{\atp{\mathrm{num}\to 0}{\mathrm{den}\to0}}\\ \lim_{x \to \infty} \log[g(x)] &=\lim_{x \to \infty}\dfrac{\frac{-\frac{3}{x^2}-\frac{10}{x^3}}{1+\frac{3}{x}+\frac{5}{x^2}}}{\frac{-1}{x^2}} =\lim_{x \to \infty}x^2\frac{\frac{3}{x^2}+\frac{10}{x^3}}{1+\frac{3}{x}+\frac{5}{x^2}}\\ \amp= \lim_{x \to \infty}\frac{3+\frac{10}{x}}{1+\frac{3}{x}+\frac{5}{x^2}} = \frac{3+0}{1+0+0}\\ \amp=3\\ \lim_{x \to \infty} f(x)&=e^3 \end{align*}Since the limit exists, \(\lim\limits_{n \to \infty} b_n =e^3\)
3.1.2.29.
Solution.
- When \(a_1=4\text{,}\) we see \(a_2=\dfrac{4+8}{3}=4\text{,}\) and so on. That is, \(a_n=4\) for every \(n\text{.}\) So, \(\lim\limits_{n \to \infty} a_n=4\text{.}\)
- Cross-multiplying, we see \(3x=x+8\text{,}\) hence \(x=4\text{.}\)
- In order for our sequence to converge to 4, the terms should be getting infinitely close to 4. So, we find the relationship between \(\textcolor{red}{a_{n+1}-4}\) and \(\textcolor{blue}{a_n-4}\text{.}\)\begin{align*} a_{n+1}&= \frac{a_n+8}{3}\\ \textcolor{red}{a_{n+1}-4}&= \frac{a_n+8}{3} -4=\frac{\textcolor{blue}{a_n-4}}{3} \end{align*}So, the distance between our sequence terms and the number 4 is decreasing by a factor of 3 each term. This implies that the terms get infinitely close to 4 as \(n\) grows. That is, \(\lim\limits_{n \to \infty}a_n=4\text{.}\)
3.1.2.30.
Solution.
- Since \(w_1\) has the highest frequency, \(w_2\) has the next-highest frequency, and so on, we know \(f_1\) is larger than the other members of its sequence, \(f_2\) is the next largest, etc. So, \(\{f_n\}\) is a decreasing sequence.
- The most-used word in a language is \(w_1\text{,}\) while the \(n\)-th most used word in a language is \(w_n\text{.}\) So, we re-state the law as:\begin{equation*} f_1=nf_n \end{equation*}Then we can rewrite this fomula a little more naturally as \(f_n=\frac{1}{n}f_1\text{.}\)
- Then \(f_3=\frac{1}{3}f_1\text{.}\) In this case, we expect the third-most used word to account for \(\frac{1}{3}(6\%) = 2\%\) of all words.
-
From (b), we know \(f_{10}=\frac{1}{10}f_1\text{.}\) Note \(f_1=6f_6=6(0.3\%)\text{.}\) Then:\begin{equation*} f_{10}=\frac{1}{10}f_1 = \frac{1}{10}6f_6=\frac{1}{10}(6)(0.3\%)=\frac{1.8}{10}\%=0.18\% \end{equation*}So, \(f_{10}\) should be \(0.18\%\) of all words.
-
The use of the word “frequency” in the statement of Zipf’s law implies \(f_n = \frac{\text{uses of $w_n$}}{\text{total number of words}}\text{.}\) The question asks for the total uses of \(w_n\text{.}\) If we call this quantity \(t_n\text{,}\) and the total number of all words is \(T\text{,}\) then Zipf’s law tells us \(\frac{t_n}{T}=\frac{1}{n}\frac{t_1}{T}\text{,}\) hence \(t_n=\frac{1}{n}t_1\text{.}\)With this notation, the problem states \(t_1=22,038,615\text{,}\) \(w_1=\texttt{the}\text{,}\) \(w_2=\texttt{be}\text{,}\) and \(w_3=\texttt{and}\text{.}\)Following Zipf’s law, \(t_n=\frac1nt_1\text{.}\) So, we expect \(t_2=\frac{t_1}{2} = 11,019,307.5\text{;}\) since this isn’t an integer, let’s say we expect \(t_2\approx 11,019,308\text{.}\) Similarly, we expect \(t_3=\frac{t_1}{3} = 7,346,205\text{.}\)Remark: The 450-million-word source material that used “the” 22,038,615 times also contained 12,545,825 instances of “be,” and 10,741,073 instances of “and.” While Zipf’s Law might be a nice model for our data overall, in these few instances it does not appear to be extremely accurate.
3.2 Series
3.2.2 Exercises
3.2.2.1.
Solution.
3.2.2.2.
Solution.
3.2.2.3.
Solution.
- We find \(\{a_n\}\) from \(\{S_N\}\) using the same logic as Question 2. \(S_N\) is the sum of the first \(N\) terms of \(\{a_n\}\text{,}\) and \(S_{N-1}\) is the sum of all the same terms except \(a_N\text{.}\) So, \(a_N = S_N-S_{N-1}\) when \(N \geq 2\text{.}\) Written another way:\begin{align*} \color{blue}{S_N}&\color{blue}{=a_1+a_2+a_3+\cdots+a_{N-2}+a_{N-1}+a_N}\\ \color{red}{S_{N-1}}&\color{red}{=a_1+a_2+a_3+\cdots+a_{N-2}+a_{N-1}}\\ \\ \end{align*}Remark: the formula given for \(S_N\) has \(S_0=0\text{,}\) which makes sense: the sum of no terms at all should be 0. However, it is common for a sequence of partial sums to start at \(N=1\text{.}\) (This fits our definition of a partial sum--we don’t really define the “sum of no terms.”) In this case, \(a_1\) must be calculated separately from the other terms of \(\{a_n\}\text{.}\) To find \(a_1\text{,}\) we simply set \(a_1=S_1\text{,}\) which (to reiterate) might not be the same as \(S_1-S_0\text{.}\)
So,
\begin{align*} \color{blue}{S_N} -\color{red}{S_{N-1}} &=\color{blue}{\Big[a_1+a_2+a_3+\cdots+a_{N-2}+a_{N-1}+a_N\Big]}\\ &- \color{red}{\Big[a_1+a_2+a_3+\cdots+a_{N-2}+a_{N-1} \Big]}\\ &=\color{blue}{a_N}\\ \end{align*}So, we calculate
\begin{align*} a_N&=S_N-S_{N-1}=\left(\dfrac{N}{N+1}\right)-\left(\dfrac{N-1}{N-1+1}\right)\\ &=\frac{N^2}{N(N+1)}-\frac{N^2-1}{N(N+1)}\\ &=\frac{1}{N(N+1)}\\ \end{align*}Therefore,
\begin{align*} a_n&=\frac{1}{n(n+1)} \end{align*} -
\begin{equation*} \lim\limits_{n \to \infty} a_n =\lim\limits_{n \to \infty} \dfrac{1}{n(n+1)}=0. \end{equation*}That is, the terms we’re adding up are getting very, very small as we go along.
- By Definition 3.2.3,\begin{equation*} \sum_{n=1}^\infty a_n = \lim_{N \to \infty}S_N = \lim_{N \to \infty}\dfrac{N}{N+1}=1 \end{equation*}That is, as we add more and more terms of our series, our cumulative sum gets very, very close to 1.
3.2.2.4.
Solution.
Note, however, that \(a_N\) is only the same as \(S_N-S_{N-1}\) when \(N \geq 2\text{:}\) otherwise, we’re trying to calculate \(S_1-S_0\text{,}\) but \(S_0\) is not defined. So, we find \(a_1\) separately:
\begin{align*} a_1&=S_1=(-1)^1+\frac{1}{1}=0\\ \end{align*}All together:
\begin{align*} a_n&=\begin{cases} 0 & \mbox{ if } n=1\\ 2(-1)^{n} -\frac{1}{n(n-1)} &\mbox{ else} \end{cases} \end{align*}3.2.2.5.
Solution.
3.2.2.6.
Solution.
- divide the top triangle into four triangles of equal area,
- colour the bottom two of them black, and
- leave the middle one white.
3.2.2.7.
Solution.
3.2.2.8.
Solution.
3.2.2.9.
Solution.
3.2.2.10.
Solution.
3.2.2.11.
Solution.
- Solution 1: Using the ideas of Question 10, we see:\begin{align*} \sum_{n={50}}^{100} \frac{1}{5^n}&=\sum_{n=0}^{100}\frac{1}{5^n}-\sum_{n=0}^{49}\frac{1}{5^n}\\ \end{align*}
That is, we want start with the sum of all the terms up to \(\dfrac{1}{5^{100}}\text{,}\) and then subtract off the ones we actually don’t want, which is everything up to \(\dfrac{1}{5^{49}}\text{.}\) Now, both series are in a form appropriate for Lemma 3.2.5.
\begin{align*} \sum_{n=0}^{100}\frac{1}{5^n}-\sum_{n=0}^{49}\frac{1}{5^n}&=\frac{1-\frac{1}{5^{101}}}{1-\frac15}-\frac{1-\frac{1}{5^{50}}}{1-\frac15}\\ &=\frac{5^{101}-1}{4\cdot 5^{100}}-\frac{5^{50}-1}{4\cdot 5^{49}}\left(\frac{5^{51}}{5^{51}}\right)\\ &=\frac{5^{101}-1}{4\cdot 5^{100}}-\frac{5^{101}-5^{51}}{4\cdot 5^{100}}\\ &=\frac{5^{51}-1}{4\cdot 5^{100}} \end{align*} - Solution 2: If we write out the first few terms of our series, we see we can factor out a constant to change the starting index.\begin{align*} \sum_{n={50}}^{100} \frac{1}{5^n}&= \frac{1}{5^{50}}+ \frac{1}{5^{51}}+\frac{1}{5^{52}}+\frac{1}{5^{53}}+\cdots +\frac{1}{5^{100}}\\ &=\frac{1}{5^{50}}\left(\frac{1}{5^0}+\frac{1}{5^1}+\frac{1}{5^2}+\frac{1}{5^3}+\cdots +\frac{1}{5^{50}}\right)\\ &=\sum_{n=0}^{50}\frac{1}{5^{50}}\cdot\frac{1}{5^n}\\ \end{align*}
Now, our sum is in the form of Lemma 3.2.5 with \(a=\dfrac{1}{5^{50}},\) \(r=\dfrac15\text{,}\) and \(N=50\text{.}\)
\begin{align*} \sum_{n=0}^{50}\frac{1}{5^{50}}\cdot\frac{1}{5^n}&=\frac{1}{5^{50}}\cdot\frac{1-\frac{1}{5^{51}}}{1-\frac15}=\frac{1-\frac{1}{5^{51}}}{4\cdot 5^{49}}\left(\frac{5^{51}}{5^{51}}\right)\\ &=\frac{5^{51}-1}{4\cdot 5^{100}} \end{align*}
3.2.2.12.
Solution.
3.2.2.13.
Solution.
So,
\begin{align*} \sum_{n=1}^\infty\left( a_n+b_n+c_{n+1}\right)&=A+B+C-\textcolor{red}{c_1} \end{align*}3.2.2.14.
Solution.
- \(\sum_{n=0}^{\infty} a_n = \sum_{n=0}^\infty b_n = \frac{1}{1-\frac12}=2\text{,}\) while
- \(\sum_{n=0}^{\infty}\dfrac{a_n}{b_n} = \sum_{n=0}^{\infty}1 =\infty \text{.}\)
3.2.2.15. (✳).
Solution.
Using Lemma 3.2.5 with \(r=\dfrac13\) and \(a=1\text{,}\)
\begin{align*} &=\frac{1}{1-\frac13}=\frac{3}{2}. \end{align*}3.2.2.16. (✳).
Solution.
- Solution 1: We write out the first few terms of the series to figure out a convenient constant to factor out.\begin{align*} \sum_{k=7}^{\infty} \frac{1}{8^k}&=\frac{1}{8^7}+\frac{1}{8^8}+\frac{1}{8^9}+\cdots\\ &=\frac{1}{8^7}\left(\frac{1}{8^0}+\frac{1}{8^1}+\frac{1}{8^2}+\cdots\right)\\ &=\sum_{k=0}^\infty \frac{1}{8^7}\cdot \frac{1}{8^n}\\ \end{align*}
We now evaluate the series using Lemma 3.2.5 with \(r=\dfrac18\text{,}\) \(a=\dfrac{1}{8^7}\text{.}\)
\begin{align*} &=\frac{1}{8^7}\cdot\dfrac{1}{1-\frac18} = \dfrac{1}{7\times 8^6} \end{align*} - Solution 2: Using the idea of Question 10, we express the series we’re interested in as the difference of two series that we can easily evaluate.\begin{align*} \sum_{k=7}^{\infty} \frac{1}{8^k}&= \sum_{k=0}^{\infty} \frac{1}{8^k}- \sum_{k=0}^{6} \frac{1}{8^k}\\ \end{align*}
Using Lemma 3.2.5,
\begin{align*} &=\frac{1}{1-\frac18}-\frac{1-\frac{1}{8^7}}{1-\frac18}\\ &=\frac{1}{7\times 8^6} \end{align*}
3.2.2.17. (✳).
Solution.
3.2.2.18. (✳).
Solution.
As \(N\to\infty\text{,}\) the argument \(\frac{\pi}{N+1}\) converges to \(0\text{,}\) and \(\cos x\) is continuous at \(x=0\text{.}\) By Definition 3.2.3, the value of the series is
\begin{align*} \lim_{N\to\infty} s_N &= \lim_{N\to\infty}\left[\frac12 - \cos\Big( \frac \pi{N+1} \Big) \bigg) \right]\\ &= \frac12 - \cos(0) = -\frac{1}{2} \end{align*}3.2.2.19. (✳).
Solution.
3.2.2.20. (✳).
Solution.
- Solution 1: If we factor our \(\left(\frac45\right)^2\text{,}\) we can change our index to something more convenient.\begin{align*} \frac{3}{2}\sum_{n=2}^\infty \left(\frac{4}{5}\right)^n&= \frac{3}{2}\sum_{n=2}^\infty \left(\frac{4}{5}\right)^2\left(\frac45\right)^{n-2}\\ &=\frac{3}{2}\sum_{n=0}^\infty \left(\frac{4}{5}\right)^2\left(\frac45\right)^{n}\\ \end{align*}
We use Lemma 3.2.5 with \(r=\dfrac45\text{.}\)
\begin{align*} &=\frac{3}{2}\left(\frac{4}{5}\right)^2\cdot\frac{1}{1-\frac45}=\frac{24}{5} \end{align*} - Solution 2: Using the idea of Question 10, we view our series as a more convenient series, minus a few initial terms.\begin{align*} \frac{3}{2}\sum_{n=2}^\infty \left(\frac{4}{5}\right)^n&= \frac{3}{2}\left(\left[\sum_{n=0}^\infty \left(\frac{4}{5}\right)^n\right] - \left(\frac{4}{5}\right)^1- \left(\frac{4}{5}\right)^0\right)\\ &= \frac{3}{2}\left(\sum_{n=0}^\infty \left(\frac{4}{5}\right)^n - \frac{9}{5}\right)\\ \end{align*}
We use Lemma 3.2.5 with \(r=\dfrac45\text{.}\)
\begin{align*} &=\frac{3}{2}\left(\frac{1}{1-\frac45} - \frac{9}{5}\right)=\frac{24}{5} \end{align*}
3.2.2.21. (✳).
Solution.
We use Lemma 3.2.5 with \(r=\dfrac{1}{10}\text{.}\)
\begin{align*} &=\frac{1}{5}+\frac{3}{10^2}\cdot\frac{1}{1-\frac{1}{10}}\\ &=\frac15+\frac{1}{30} = \frac{7}{30} \end{align*}3.2.2.22. (✳).
Solution.
We use Lemma 3.2.5 with \(r=\dfrac{1}{100}\text{.}\)
\begin{align*} &= 2 + \frac{65}{100}\cdot\frac{1}{1-\frac{1}{100}}\\ &=2+\frac{65}{99} = \frac{263}{99} \end{align*}3.2.2.23. (✳).
Solution.
We use Lemma 3.2.5 with \(r=\dfrac{1}{10^3}\text{.}\)
\begin{align*} &=\frac{321}{1000} \cdot\frac{1}{1-\frac{1}{10^3}}=\frac{321}{999}=\frac{107}{333} \end{align*}3.2.2.24. (✳).
Solution.
We use Lemma 3.2.5 with \(r=\dfrac{2}{3}\) and \(a=\dfrac{8}{9}\text{.}\)
\begin{align*} &=\frac{8}{9}\cdot\frac{1}{1-\frac23}=\color{blue}{\frac{8}{3}} \end{align*}because all the terms except the first part of the \(n=2\) term, and the last part of the \(n=N\) term, cancel. Then:
\begin{align*} \color{red}{\sum_{n=2}^\infty \bigg(\frac1{2n-1} - \frac1{2n+1}\bigg)} &=\lim_{N \to \infty}s_N = \lim_{N \to \infty}\frac{1}{3}-\frac{1}{2N+1}\\ &=\color{red}{\frac{1}{3}} \end{align*}3.2.2.25. (✳).
Solution.
Both are geometric series.
\begin{align*} &= \sum_{n=0}^\infty {\Big(\frac{1}{3}\Big)}^{n+1} + \sum_{n=0}^\infty {\Big(-\frac{2}{5}\Big)}^{n}\\ &= \frac{1}{3}\sum_{n=0}^\infty {\Big(\frac{1}{3}\Big)}^{n} + \sum_{n=0}^\infty {\Big(-\frac{2}{5}\Big)}^{n}\\ \end{align*}We use Lemma 3.2.5 with \(a_1=\dfrac{1}{3}\) and \(r_1=\dfrac{1}{3}\text{,}\) then with \(a_2=1\) and \(r_2=-\dfrac{2}{5}\text{.}\)
\begin{align*} &=\frac{1}{3}\cdot\frac{1}{1-\frac13}+\frac{1}{1+\frac25}\\ &=\frac{1}{2}+\frac{5}{7}=\frac{17}{14} \end{align*}3.2.2.26. (✳).
Solution.
Using Lemma 3.2.5,
\begin{align*} &=\frac{1}{1-\frac14}+\frac{3}{1-\frac{3}{4}}\\ &=\frac{4}{3}+12=\frac{40}{3} \end{align*}3.2.2.27.
Solution.
when \(N \geq 7\text{.}\) So,
\begin{align*} \sum_{n=5}^\infty \Big(\log(n-3)-\log(n)\Big)&= \lim_{N \to \infty}s_N\\ &=\lim_{N \to \infty}\log \left(\frac{24}{N(N-1)(N-2)}\right)\\ &=-\infty \end{align*}3.2.2.28.
Solution.
3.2.2.29.
Solution.
3.2.2.30.
Solution.
So, the volume of all the spheres together is:
\begin{align*} \sum_{n=1}^\infty v_n&=\sum_{n=1}^\infty \frac{4\pi}{3}\left(\frac{1}{\pi^3}\right)^{n}\\ &=\sum_{n=0}^\infty \frac{4\pi}{3}\left(\frac{1}{\pi^3}\right)^{n+1}\\ &=\sum_{n=0}^\infty \frac{4}{3\pi^2}\left(\frac{1}{\pi^3}\right)^{n}\\ \end{align*}We use Lemma 3.2.5 with \(a= \dfrac{4}{3\pi^2}\) and \(r=\dfrac{1}{\pi^3}\text{.}\)
\begin{align*} &= \dfrac{4}{3\pi^2}\cdot\frac{1}{1-\frac{1}{\pi^3}} = \frac{4\pi}{3\left(\pi^3-1\right)} \end{align*}3.2.2.31.
Solution.
Using Definition 3.2.3, our series evaluates to:
\begin{align*} \lim_{N \to \infty}s_N &=\lim_{N \to \infty}\left[\frac{\sin^23}{2^3} +\left( \sum_{n=4}^{N}\frac{1}{2^n} \right)+\frac{\cos^2(N+1)}{2^{N+1}}\right]\\ &=\frac{\sin^23}{8}+\left[\lim_{N \to \infty}\frac{\cos^2(N+1)}{2^{N+1}}\right] +\sum_{n=4}^\infty \frac{1}{2^n}\\ \end{align*}We evaluate the limit using the squeeze theorem; the series is geometric.
\begin{align*} &=\frac{\sin^23}{8}+0+\sum_{n=0}^\infty \frac{1}{2^{n+4}}\\ &=\frac{\sin^23}{8}+\frac{1}{2^4}\sum_{n=0}^\infty \frac{1}{2^{n}}\\ \end{align*}Using Lemma 3.2.5,
\begin{align*} &=\frac{\sin^23}{8}+\frac{1}{2^4} \frac{1}{1-\frac12}\\ &=\frac{\sin^23}{8}+\frac{1}{8} \approx 0.1275 \end{align*}3.2.2.32.
Solution.
Similarly, we find \(\{a_n\}\) from \(\{S_N\}\text{.}\) Do be careful: \(S_N\) only follows the formula we found above when \(N \geq 2\text{.}\) In the next line, we use an expression containing \(S_{n-1}\text{;}\) in order for the subscript to be at least two (so the formula fits), we need \(n \geq 3\text{.}\)
\begin{align*} a_n&=S_n-S_{n-1}=-\frac{1}{n(n-1)}-\frac{-1}{(n-1)(n-2)}\\ &=\frac{2}{n(n-1)(n-2)} &\text{if }n \ge 3,\\ a_2&=S_2-S_1=-\frac{1}{2(2-1)} - 2=-\frac52,\\ a_1&=S_1=2\\ \end{align*}All together,
\begin{align*} a_n &=\begin{cases} \frac{2}{n(n-1)(n-2)} &\text{ if }n \ge 3,\\ -\frac52 &\text{ if }n=2,\\ 2 &\text{ if }n=1 \end{cases} \end{align*}3.2.2.33.
Solution.
3.3 Convergence Tests
3.3.11 Exercises
3.3.11.1.
Solution.
- \(\displaystyle\lim_{n \to \infty}\frac{1}{n}=0\text{,}\) so the divergence test is inconclusive. It’s true that this series diverges, but we can’t show it using the divergence test.
- \(\displaystyle\lim_{n \to \infty}\frac{n^2}{n+1}=\infty\text{,}\) which is not zero, so the divergence test tells us this series diverges.
-
We’ll show below that \(\displaystyle\lim_{n \to \infty}\sin n\) does not exist. In particular, it is not zero. Therefore, the divergence test tells us this series diverges.Now we’ll show that \(\displaystyle\lim_{n \to \infty}\sin n\) does not exist. Suppose that it does exist and takes the value \(S\text{.}\) We will now see that this assumption leads to a contradiction. Add together the two trig identities (see Appendix A.8)\begin{align*} \sin(n+1) &= \sin(n)\cos(1)+\cos(n)\sin(1) \\ \sin(n-1) &= \sin(n)\cos(1)-\cos(n)\sin(1) \end{align*}This gives\begin{equation*} \sin(n+1) +\sin (n-1) = 2\sin(n)\cos(1) \end{equation*}Taking the limit \(n\rightarrow\infty\) gives \(2S=2S\cos(1)\text{.}\) Since \(\cos(1)\ne 1\text{,}\) this forces \(S=0\text{.}\) Now the first trig identity above gives\begin{equation*} \cos(n) = \frac{\sin(n+1)-\sin(n)\cos(1)}{\sin(1)} \end{equation*}Taking the limit as \(n\rightarrow\infty\) of that gives\begin{equation*} \lim_{n\to\infty}\cos(n) = \frac{S-S\cos(1)}{\sin(1)}=0 \end{equation*}But that provides the contradiction. Because \(\sin^2(n)+\cos^2(n)=1\text{,}\) we can’t have both \(\sin(n)\) and \(\cos(n)\) converging to zero. So \(\displaystyle\lim_{n \to \infty}\sin n\) does not exist.
- For all whole numbers \(n\text{,}\) \(\sin(\pi n)=0\text{,}\) so \(\displaystyle\lim_{n \to \infty}\sin(\pi n)=0\) and the divergence test is inconclusive.
3.3.11.2.
Solution.
- \(f(x) = \frac{1}{x}\text{,}\) which is positive and decreasing for all \(x \ge 1\text{,}\) so the integral test does apply here.
- \(f(x)=\frac{x^2}{x+1}\text{,}\) which is not decreasing — in fact, it goes to infinity. So, the integral test does not apply here. (The divergence test tells us the series diverges, though.)
- \(f(x) = \sin x\text{,}\) which is neither consistently positive nor consistently decreasing, so the integral test does not apply. (The divergence test tells us the series diverges, though.)
- \(f(x)=\frac{\sin x+1}{x^2}\) is positive for all whole numbers \(n\text{.}\) To determine whether it is decreasing, we consider its derivative.\begin{align*} f'(x)&=\frac{x^2(\cos x)-(\sin x+1)(2x)}{x^4}=\frac{x\cos x - 2\sin x -2}{x^3} \end{align*}This is sometimes positive, and sometimes negative. (For example, if \(x=100\pi\text{,}\) \(f'(x) = \frac{100\pi-0-2}{(100\pi)^3} \gt 0\text{,}\) but if \(x=101\pi\) then \(f'(x)=\frac{101\pi(-1)-0-2}{(101\pi)^3} \lt 0\text{.}\)) Then \(f(x)\) is not a decreasing function, so the integral test does not apply.
3.3.11.3.
Solution.
3.3.11.4.
Solution.
if \(\sum a_n\) converges | if \(\sum a_n\) diverges | |
and if \(\{a_n\}\) is the red series | then \(\sum b_n\) CONVERGES | inconclusive |
and if \(\{a_n\}\) is the blue series | inconclusive | then \(\sum b_n\) DIVERGES |
3.3.11.5.
Solution.
3.3.11.6.
Solution.
3.3.11.7.
Solution.
3.3.11.8.
Solution.
3.3.11.9.
Solution.
3.3.11.10.
Solution.
3.3.11.11.
Solution.
3.3.11.12.
Solution.
3.3.11.13.
Solution.
- \(\displaystyle\int_{1}^\infty \sin(\pi x)\,\dee{x} = \lim_{R \to \infty} \left[-\frac{1}{\pi}\cos (\pi x)\right]_1^R = \lim_{R \to \infty}\Big[-\cos (\pi R)\Big]-\frac{1}{\pi}\) Since the limit does not exist, the integral diverges.
- \(\displaystyle\sum_{n=1}^{\infty}\sin(\pi n) = \sum_{n=1}^\infty 0 = 0\text{.}\) The series converges.
3.3.11.14. (✳).
Solution.
3.3.11.15. (✳).
Solution.
3.3.11.16. (✳).
Solution.
3.3.11.17. (✳).
Solution.
3.3.11.18. (✳).
Solution.
3.3.11.19.
Solution.
3.3.11.20.
Solution.
3.3.11.21.
Solution.
3.3.11.22.
Solution.
3.3.11.23.
Solution.
3.3.11.24.
Solution.
Since \(e\) is a constant,
\begin{align*} \lim_{k \rightarrow \infty}\frac{a_{k+1}}{a_k}&=\lim_{k \rightarrow \infty}\frac{e}{k+1}=0 \end{align*}3.3.11.25.
Solution.
Now it looks like a geometric series with \(r=\frac{2}{3}\)
\begin{align*} &=\frac{1}{9}\left(\frac{1}{1-(2/3)}\right)=\frac{1}{3} \end{align*}3.3.11.26.
Solution.
3.3.11.27.
Solution.
3.3.11.28. (✳).
Solution.
3.3.11.29. (✳).
Solution.
3.3.11.30. (✳).
Solution.
3.3.11.31. (✳).
Solution.
3.3.11.32. (✳).
Solution.
3.3.11.33. (✳).
Solution.
-
Solution 1: Let’s see whether the divergence test works here.\begin{equation*} \lim_{n\to\infty} \frac{n^4 2^{n/3}}{(2n+7)^4}\left(\frac{\frac{1}{n^4}}{\frac{1}{n^4}}\right) = \lim_{n\to\infty} \frac{2^{n/3}}{(2+7/n)^4} = \lim_{n\to\infty} \frac{2^{n/3}}{(2+0)^4} = \infty \end{equation*}The summands of our series do not converge to zero. By the divergence test, the series diverges.
-
Solution 2: Let’s develop some intuition for a comparison. For very large \(n\text{,}\) \(2n\) dominates \(7\) so that\begin{gather*} \frac{n^4 2^{n/3}}{(2n+7)^4} \approx \frac{n^4 2^{n/3}}{(2n)^4} =\frac{1}{16}2^{n/3} \end{gather*}The series \(\displaystyle \sum_{n=1}^{\infty} 2^{n/3}\) is a geometric series with ratio \(r=2^{1/3} \gt 1\) and so diverges. (It also fails the divergence test.) We expect the given series to diverge too.To verify that our intuition is correct, we apply the limit comparison test with\begin{gather*} a_n= \frac{n^4 2^{n/3}}{(2n+7)^4} \quad\text{and}\quad b_n= 2^{n/3} \end{gather*}which is valid since\begin{equation*} \lim_{n\rightarrow\infty} \frac{a_n}{b_n} =\lim_{n\rightarrow\infty}\frac{n^4}{(2n+7)^4} =\lim_{n\rightarrow\infty}\frac{1}{{(2+7/n)}^4} =\frac{1}{2^4} \end{equation*}exists and is nonzero. Since the series \(\sum\limits_{n=1}^\infty b_n\) is a divergent geometric series (with ratio \(r=2^{1/3} \gt 1\)), the given series diverges.(It is possible to use the plain comparison test as well. One needs to show something like \(a_n = \frac{n^4 2^{n/3}}{(2n+7)^4} \ge \frac{n^4 2^{n/3}}{(2n+7n)^4} = \frac{1}{9^4}b_n\text{.}\))
- Solution 3: Alternately, one can apply the ratio test:\begin{align*} \lim_{n\to\infty} \bigg| \frac{a_{n+1}}{a_n} \bigg| &= \lim_{n\to\infty} \bigg| \frac{(n+1)^42^{(n+1)/3}/(2(n+1)+7)^4}{n^42^{n/3}/(2n+7)^4} \bigg|\\ &= \lim_{n\to\infty} \frac{(n+1)^4(2n+7)^4}{n^4(2n+9)^4}\ \frac{2^{(n+1)/3}}{2^{n/3}}\\ &=\lim_{n\to\infty} \frac{(1+1/n)^4(2+7/n)^4}{(2+9/n)^4} \cdot2^{1/3} = 1\cdot2^{1/3} \gt 1. \end{align*}Since the ratio of consecutive terms is greater than one, by the ratio test, the series diverges.
3.3.11.34. (✳).
Solution.
3.3.11.35. (✳).
Solution.
3.3.11.36. (✳).
Solution.
3.3.11.37.
Solution.
3.3.11.38. (✳).
Solution.
We use Lemma 3.2.5 with \(a=\frac{6}{7^2}\) and \(r=\frac{1}{7}\text{.}\)
\begin{align*} &=\frac{6}{7^2}\cdot\frac{1}{1-\frac17}=\frac{6}{42}=\frac17 \end{align*}3.3.11.39. (✳).
Solution.
- Solution 1: The given series is\begin{gather*} 1+\frac{1}{3}+\frac{1}{5}+\frac{1}{7}+\frac{1}{9}+\cdots =\sum_{n=1}^\infty a_n \text{ with } \end{gather*}First we’ll develop some intuition by observing that, for very large \(n\text{,}\) \(a_n\approx \frac{1}{2n}\text{.}\) We know that the series \(\sum\limits_{n=1}^\infty\frac{1}{n}\) diverges by the \(p\)-test with \(p=1\text{.}\) So let’s apply the limit comparison test with \(b_n=\frac{1}{n}\text{.}\) Since\begin{gather*} \lim_{n\rightarrow\infty}\frac{a_n}{b_n} =\lim_{n\rightarrow\infty}\frac{n}{2n-1} =\lim_{n\rightarrow\infty}\frac{1}{2-\frac{1}{n}} =\frac{1}{2} \end{gather*}the series \(\sum\limits_{n=1}^\infty a_n\) converges if and only if the series \(\sum\limits_{n=1}^\infty b_n\) converges. So the given series diverges.
- Solution 2: The series\begin{align*} 1+\frac{1}{3}+\frac{1}{5}+\frac{1}{7}+\frac{1}{9}+\cdots &\ge \frac{1}{2}+\frac{1}{4}+\frac{1}{6}+\frac{1}{8}+\frac{1}{10}+\cdots\\ &=\frac{1}{2}\Big(1+\frac{1}{2}+\frac{1}{3}+\frac{1}{4}+\frac{1}{5}+\cdots\Big) \end{align*}The series in the brackets is the harmonic series which we know diverges, by the \(p\)-test with \(p=1\text{.}\) So the series on the right hand side diverges. By the direct comparison test, the series on the left hand side diverges too.
3.3.11.40. (✳).
Solution.
3.3.11.41. (✳).
Solution.
3.3.11.42. (✳).
Solution.
3.3.11.43. (✳).
Solution.
3.3.11.44. (✳).
Solution.
3.3.11.45. (✳).
Solution.
3.3.11.46. (✳).
Solution.
- Solution 1:
- Our first task is to identify the potential sources of impropriety for this integral.
- The domain of integration extends to \(+\infty\text{.}\) On the domain of integration the denominator is never zero so the integrand is continuous. Thus the only problem is at \(+\infty\text{.}\)
- Our second task is to develop some intuition about the behavior of the integrand for very large \(x\text{.}\) When \(x\) is very large:
- \(|\sin x|\le 1 \ll x\text{,}\) so that the numerator \(x+\sin x\approx x\text{,}\) and
- \(1 \ll x^2\text{,}\) so that denominator \(1+x^2\approx x^2\text{,}\) and
- the integrand \(\displaystyle\frac{x+\sin x}{1+x^2} \approx \frac{x}{x^2} =\frac{1}{x}\)
- Now, since \(\displaystyle\int_2^\infty\frac{\dee{x}}{x}\) diverges, we would expect \(\displaystyle\int_2^\infty\frac{x+\sin x}{1+x^2}\ \dee{x}\) to diverge too.
- Our final task is to verify that our intuition is correct. To do so, we set\begin{align*} f(x) &= \frac{x+\sin x}{1+x^2} & g(x) &= \frac{1}{x} \end{align*}and compute\begin{align*} \lim_{x\rightarrow\infty}\frac{f(x)}{g(x)} &=\lim_{x\rightarrow\infty} \frac{x+\sin x}{1+x^2}\div\frac{1}{x}\\ &=\lim_{x\rightarrow\infty} \frac{(1+\sin x/x)x}{(1/x^2+1)x^2}\times x\\ &=\lim_{x\rightarrow\infty} \frac{1+\sin x/x}{1/x^2+1}\\ &=1 \end{align*}
- Solution 2: Let’s break up the integrand as \(\dfrac{x+\sin x}{1+x^2} = \dfrac{x}{1+x^2}+\dfrac{\sin x}{1+x^2}\text{.}\) First, we consider the integral \(\displaystyle \int_2^\infty\frac{\sin x}{1+x^2}\ \dee{x}\text{.}\)
- \(\displaystyle\frac{|\sin x|}{1+x^2}\le \frac{1}{1+x^2}\text{,}\) so if we can show \(\displaystyle\int \frac{1}{1+x^2}\,\dee{x}\) converges, we can conclude that \(\displaystyle\int\frac{|\sin x|}{1+x^2}\,\dee{x}\) converges as well by the comparison test.
- \(\displaystyle \displaystyle\int_2^\infty\frac{1}{1+x^2}\ \dee{x}\le \int_2^\infty\frac{1}{x^2}\ \dee{x}\)
- \(\displaystyle\ \int_2^\infty\frac{1}{x^2}\ \dee{x}\) converges (by the \(p\)-test with \(p=2\))
- So the integral \(\displaystyle \int_2^\infty\frac{\sin x}{1+x^2}\ \dee{x}\) converges by the comparison test, and hence
- \(\displaystyle \int_2^\infty\frac{\sin x}{1+x^2}\ \dee{x}\) converges as well.
Therefore, \(\displaystyle \int_2^\infty\frac{x+\sin x}{1+x^2}\ \dee{x}\) converges if and only if \(\displaystyle \int_2^\infty\frac{x}{1+x^2}\ \dee{x}\) converges. But\begin{gather*} \int_2^\infty\frac{x}{1\!+\!x^2}\ \dee{x} =\lim_{r\rightarrow\infty}\int_2^r\frac{x}{1\!+\!x^2}\ \dee{x} =\lim_{r\rightarrow\infty}\Big[\half\log(1\!+\!x^2)\Big]_2^r=\infty \end{gather*}diverges, so \(\displaystyle\int_2^\infty\frac{x+\sin x}{1+x^2}\ \dee{x}\) diverges.
- Solution 1: Set \(a_n= \frac{n+\sin n}{1+n^2}\text{.}\) We first try to develop some intuition about the behaviour of \(a_n\) for large \(n\) and then we confirm that our intuition was correct.
- Step 1: Develop intuition. When \(n\gg 1\text{,}\) the numerator \(n+\sin n\approx n\text{,}\) and the denominator \(1+n^2\approx n^2\) so that \(a_n\approx \frac{n}{n^2}=\frac{1}{n}\) and it looks like our series should diverge by the \(p\)-test (Example 3.3.6) with \(p=1\text{.}\)
- Step 2: Verify intuition. To confirm our intuition we set \(b_n=\frac{1}{n}\) and compute the limit\begin{align*} \lim_{n\rightarrow\infty}\frac{a_n}{b_n} \amp=\lim_{n\rightarrow\infty}\frac{ \frac{n+\sin n}{1+n^2} } {\frac{1}{n}} =\lim_{n\rightarrow\infty}\frac{n[n+\sin n]} {1+n^2}\\ \amp=\lim_{n\rightarrow\infty}\frac{1+\frac{\sin n}{n}} {\frac{1}{n^2}+1} =1 \end{align*}We already know that the series \(\sum\limits_{n=1}^\infty b_n =\sum\limits_{n=1}^\infty\frac{1}{n}\) diverges by the \(p\)-test with \(p=1\text{.}\) So our series diverges by the limit comparison test, Theorem 3.3.11.
- Solution 2: Since \(\big|\frac{\sin n}{1+n^2}\big|\le\frac{1}{n^2}\) and the series \(\sum\limits_{n=1}^\infty \frac{1}{n^2}\) converges by the \(p\)-test with \(p=2\text{,}\) the series \(\sum\limits_{n=1}^\infty \frac{\sin n}{1+n^2}\) converges. Hence \(\sum\limits_{n=1}^\infty \frac{n+\sin n}{1+n^2}\) converges if and only if the series \(\sum\limits_{n=1}^\infty \frac{n}{1+n^2}\) converges. Now \(f(x)=\frac{x}{1+x^2}\) is a continuous, positive, decreasing function on \([1,\infty)\) since\begin{gather*} f'(x)=\frac{(1+x^2)-x(2x)}{{(1+x^2)}^2} =\frac{1-x^2}{{(1+x^2)}^2} \end{gather*}is negative for all \(x \gt 1\text{.}\) We saw in part (a) that the integral \(\int_2^\infty\frac{x}{1+x^2}\ \dee{x}\) diverges. So the integral \(\int_1^\infty\frac{x}{1+x^2}\ \dee{x}\) diverges too and the sum \(\sum\limits_{n=1}^\infty \frac{n}{1+n^2}\) diverges by the integral test. So \(\sum\limits_{n=1}^\infty \frac{n+\sin n}{1+n^2}\) diverges.
3.3.11.47. (✳).
Solution.
3.3.11.48. (✳).
Solution.
24
From the information in the problem statement, we know
\begin{align*} \sum_{n=N+1}^\infty 2a_n&=2\sum_{n=N+1}^\infty a_n\qquad\text{converges. }\\ \end{align*}So, by the direct comparison test,
\begin{align*} \sum_{n=N+1}^\infty\frac{a_n}{1-a_n}&\qquad\text{converges as well. }\\ \end{align*}Since the convergence of a series is not affected by its first \(N\) terms, as long as \(N\) is finite, we conclude
\begin{align*} \sum_{n=1}^\infty\frac{a_n}{1-a_n}&\qquad\text{converges.} \end{align*}3.3.11.49. (✳).
Solution.
3.3.11.50. (✳).
Solution.
3.3.11.51. (✳).
Solution.
By the direct comparison test,
\begin{align*} \sum_{n=N+1}^\infty a_n^2& \qquad\text{converges.}\\ \end{align*}Since convergence doesn’t depend on the first \(N\) terms of a series for any finite \(N\text{,}\)
\begin{align*} \sum_{n=1}^\infty a_n^2& \qquad\text{converges as well.} \end{align*}3.3.11.52.
Solution.
3.3.11.53.
Solution.
3.4 Absolute and Conditional Convergence
3.4.3 Exercises
3.4.3.1. (✳).
Solution.
3.4.3.2.
Solution.
\(\sum a_n\) converges | \(\sum a_n\) diverges | |
\(\sum |a_n|\) converges | converges absolutely | not possible |
\(\sum |a_n|\) diverges | converges conditionally | diverges |
3.4.3.3. (✳).
Solution.
3.4.3.4. (✳).
Solution.
3.4.3.5. (✳).
Solution.
3.4.3.6. (✳).
Solution.
- Step 1: Since \(|\cos n|\le 1\text{,}\) we have\begin{equation*} \left|\frac{\sqrt{n}\cos(n)}{n^2-1}\right| \le \frac{\sqrt{n}}{n^2-1} \end{equation*}for all \(n\gt 1\text{.}\) So, by part (a) of the comparison test, which is Theorem 3.3.8, if the series \(\displaystyle \sum_{n=5}^\infty \frac{\sqrt{n}}{n^2-1}\) converges, then we will have that the series \(\displaystyle \sum_{n=5}^\infty \left|\frac{\sqrt{n}\cos(n)}{n^2-1}\right|\) also converges, and hence that the series \(\displaystyle \sum_{n=5}^\infty \frac{\sqrt{n}\cos(n)}{n^2-1}\) converges absolutely.
- Step 2: Now, to prove that the series \(\displaystyle \sum_{n=5}^\infty \frac{\sqrt{n}}{n^2-1}\) converges, we apply the limit comparison test with \(a_n=\frac{\sqrt{n}}{n^2-1}\) and \(b_n = \frac{1}{n^{3/2}}\) (for \(n\ge 5\)). Since\begin{align*} \lim_{n \to \infty}\frac{a_n}{b_n} &=\lim_{n \to \infty} \frac{\frac{\sqrt{n}}{n^2-1}}{\frac{1}{n^{3/2}}} = \lim_{n \to \infty}\frac{\sqrt{n}\cdot\sqrt{n}^3}{n^2-1}\\ &=\lim_{n \to \infty}\frac{n^2}{n^2-1} = \lim_{n \to \infty}\frac{1}{1-1/n^2}\\ & = 1 \end{align*}and since \(\displaystyle \sum_{n=5}^\infty \frac{1}{n^{3/2}}\) converges by the \(p\)-test, the limit comparison test tells us that the series \(\displaystyle\sum_{n=5}^\infty \frac{\sqrt{n}}{n^2-1}\) converges. So, by step 1, \(\displaystyle\sum_{n=5}^\infty \frac{\sqrt{n}\cos(n)}{n^2-1}\) converges absolutely.
3.4.3.7. (✳).
Solution.
3.4.3.8. (✳).
Solution.
3.4.3.9. (✳).
Solution.
3.4.3.10.
Solution.
3.4.3.11.
Solution.
3.4.3.12.
Solution.
3.4.3.13. (✳).
Solution.
- Solution 1: We need to show that \(\sum\limits_{n=1}^\infty 24n^2 e^{-n^3}\) converges. If we replace \(n\) by \(x\) in the summand, we get \(f(x) = 24x^2 e^{-x^3}\text{,}\) which we can integate. (Just substitute \(u=x^3\text{.}\)) So let’s try the integral test. First, we have to check that \(f(x)\) is positive and decreasing. It is certainly positive. To determine if it is decreasing, we compute\begin{gather*} \diff{f}{x} = 48x e^{-x^3} - 24\times 3 x^4 e^{-x^3} = 24x (2-3x^3) e^{-x^3} \end{gather*}which is negative for \(x\ge1\text{.}\) Therefore \(f(x)\) is decreasing for \(x\ge1\text{,}\) and the integral test applies. The substitution \(u=x^3\text{,}\) \(\dee{u}=3x^2\,\,\dee{x}\text{,}\) yields\begin{align*} \int f(x) \,\dee{x} \amp= \int 24x^2 e^{-x^3} \,\dee{x} = \int 8 e^{-u}\,\dee{u} = -8e^{-u} + C \\ \amp= -8e^{-x^3} + C. \end{align*}Therefore\begin{align*} \int_1^\infty f(x) \,\dee{x} &= \lim_{R \to \infty} \int_1^R f(x) \,\dee{x} = \lim_{R \to \infty} \bigg[ {-}8 e^{-x^3} \bigg]_1^R\\ &= \lim_{R \to \infty} ( - 8 e^{-R^3} + 8 e^{-1} ) = 8 e^{-1} \end{align*}Since the integral is convergent, the series \(\sum\limits_{n=1}^\infty 24n^2 e^{-n^3}\) converges and the series \(\displaystyle \sum_{n=1}^\infty (-1)^{n-1}24n^2 e^{-n^3}\) converges absolutely.
- Solution 2: Alternatively, we can use the ratio test with \(a_n=24n^2 e^{-n^3}\text{.}\) We calculate\begin{align*} \lim_{n\to\infty} \left|\frac{a_{n+1}}{a_n}\right| &= \lim_{n\to\infty} \left|\frac{24 (n+1)^2e^{-(n+1)^3}}{24 n^2e^{-n^3}} \right|\\ &= \lim_{n\to\infty} \left( \frac{(n+1)^2}{n^2} \frac{e^{n^3}}{e^{(n+1)^3}} \right)\\ &= \lim_{n\to\infty} \left(1+\frac{1}{n}\right)^2 e^{-(3n^2+3n+1)} = 1\cdot0=0 \lt 1, \end{align*}and therefore the series converges absolutely.
- Solution 3: Alternatively, alternatively, we can use the limiting comparison test. First a little intuition building. Recall that we need to show that \(\sum\limits_{n=1}^\infty 24n^2 e^{-n^3}\) converges. The \(n^{\rm th}\) term in this series is\begin{equation*} a_n = 24n^2 e^{-n^3} =\frac{24 n^2}{e^{n^3}} \end{equation*}It is a ratio with both the numerator and denominator growing with \(n\text{.}\) A good rule of thumb is that exponentials grow a lot faster than powers. For example, if \(n=10\) the numerator is \(2400=2.4\times 10^3\) and the denominator is about \(2\times 10^{434}\text{.}\) So we would guess that \(a_n\) tends to zero as \(n\rightarrow\infty\text{.}\) The question is “does \(a_n\) tend to zero fast enough with \(n\) that our series converges?”. For example, we know that \(\sum_{n=1}^\infty \frac{1}{n^2}\) converges (by the \(p\)-test with \(p=2\)). So if \(a_n\) tends to zero faster than \(\frac{1}{n^2}\) does, our series will converge. So let’s try the limiting convergence test with \(a_n = 24n^2 e^{-n^3} =\frac{24 n^2}{e^{n^3}}\) and \(b_n=\frac{1}{n^2}\text{.}\)\begin{gather*} \lim_{n\rightarrow\infty}\frac{a_n}{b_n} =\lim_{n\rightarrow\infty}\frac{24n^2 e^{-n^3}}{1/n^2} =\lim_{n\rightarrow\infty}\frac{24n^4 }{e^{n^3}} \end{gather*}By l’Hôpital’s rule, twice,\begin{align*} \lim_{x\rightarrow\infty} \frac{24x^4 }{e^{x^3}} &=\lim_{x\rightarrow\infty} \frac{4\times 24x^3 }{3 x^2e^{x^3}} &\text{by l'Hôpital}\\ &=\lim_{x\rightarrow\infty} \frac{32x }{e^{x^3}} &\text{just cleaning up}\\ &=\lim_{x\rightarrow\infty} \frac{32 }{3x^2e^{x^3}} &\text{by l'Hôpital, again}\\ &=0 \end{align*}That’s it. The limit comparison test now tells us that \(\sum_{n=1}^\infty a_n\) converges.
3.4.3.14.
Solution.
3.4.3.15.
Solution.
3.5 Power Series
3.5.3 Exercises
3.5.3.1.
Solution.
This is a geometric series with \(r=\frac{1}{2}\text{,}\) so we know that it converges and
\begin{align*} &=\frac{1}{1-\frac12}=2 \end{align*}3.5.3.2.
Solution.
3.5.3.3.
Solution.
3.5.3.4.
Solution.
3.5.3.5. (✳).
Solution.
3.5.3.6. (✳).
Solution.
3.5.3.7. (✳).
Solution.
3.5.3.8. (✳).
Solution.
3.5.3.9. (✳).
Solution.
3.5.3.10. (✳).
Solution.
3.5.3.11. (✳).
Solution.
3.5.3.12. (✳).
Solution.
3.5.3.13. (✳).
Solution.
3.5.3.14. (✳).
Solution.
3.5.3.15. (✳).
Solution.
3.5.3.16. (✳).
Solution.
3.5.3.17.
Solution.
- Solution 1: Using the Fundamental Theorem of Calculus Part 1:\begin{align*} \diff{}{x} \left\{\int_5^x f(t)\dee{t} \right\}&=f(x)\\ \text{So,}\qquad f(x)&=\diff{}{x} \left\{3x+\displaystyle\sum_{n=0}^\infty \frac{(x-1)^{n+1}}{n(n+1)^2} \right\}\\ &=3+\displaystyle\sum_{n=1}^\infty \frac{(n+1)(x-1)^{n}}{n(n+1)^2}\\ &=3+\displaystyle\sum_{n=1}^\infty \frac{(x-1)^{n}}{n(n+1)} \end{align*}
-
Solution 2: Suppose we had used \(f'(x)\) instead. We would antidifferentiate to find:\begin{align*} f(x)&=\int \left( \sum_{n=0}^\infty \frac{(x-1)^{n}}{n+2} \right)\dee{x}\\ &=\left(\sum_{n=0}^\infty \frac{(x-1)^{n+1}}{(n+1)(n+2)}\right)+C\\ &=\left(\sum_{n=1}^\infty \frac{(x-1)^{n}}{n(n+1)}\right)+C \end{align*}Notice \(f(1)=0+C\text{.}\) So, to find \(C\text{,}\) we must find \(f(1)\text{.}\) We can’t get that information from \(f'(x)\text{,}\) so our only option is to consider the given formula for \(\int_5^x f(t)\dee{t}\text{.}\) Using the Fundamental Theorem of Calculus Part 1:\begin{align*} f(1)&=\left.\diff{}{x}\left\{\int_5^x f(t)\dee{t} \right\}\right|_{x=1}\\ &=\left.\diff{}{x}\left\{ 3x+\displaystyle\sum_{n=1}^\infty \frac{(x-1)^{n+1}}{n(n+1)^2} \right\}\right|_{x=1}\\ &=\left[ 3+\displaystyle\sum_{n=1}^\infty \frac{(n+1)(x-1)^{n}}{n(n+1)^2} \right]_{x=1}\\ &=\left[ 3+\displaystyle\sum_{n=1}^\infty \frac{(x-1)^{n}}{n(n+1)} \right]_{x=1}\\ &= 3+\displaystyle\sum_{n=1}^\infty \frac{0^{n}}{n(n+1)}\\ &=3 \end{align*}So, \(f(x)=3+\displaystyle\sum_{n=1}^\infty \frac{(x-1)^{n}}{n(n+1)}\text{.}\)Note that in Solution 2, we did the same calculation as Solution 1, and more.
3.5.3.18. (✳).
Solution.
- once for the series whose \(n^{\rm th}\) term is \(a_n=\frac{x^n}{3^{2n}\log n}\text{,}\) and
- once for the series whose \(n^{\rm th}\) term is \(a_n=\big|\frac{x^n}{3^{2n}\log n}\big|\text{.}\)
3.5.3.19. (✳).
Solution.
3.5.3.20. (✳).
Solution.
3.5.3.21. (✳).
Solution.
3.5.3.22. (✳).
Solution.
3.5.3.23.
Solution.
Differentiating both sides with respect to \(x\text{:}\)
\begin{align*} \sum_{n=0}^\infty nx^{n-1}&=\frac{1}{(1-x)^2}\\ \sum_{n=1}^\infty nx^{n-1}&=\frac{1}{(1-x)^2}\\ \end{align*}Multiplying both sides by \(x\text{:}\)
\begin{align*} \sum_{n=1}^\infty nx^{n}&=\frac{x}{(1-x)^2} \end{align*}Therefore,
\begin{align*} \bar x &=\frac{5/4}{3/2}=\frac{5}{6} = 0.8\overline{33} \end{align*}3.5.3.24.
Solution.
3.5.3.25.
Solution.
3.5.3.26.
Solution.
3.5.3.27.
Solution.
3.6 Taylor Series
3.6.8 Exercises
3.6.8.1.
Solution.
3.6.8.2.
Solution.
3.6.8.3.
Solution.
- The Taylor series representation of \(\frac{1}{1-x}\) is given in V. The series converges for \(-1\lt x\lt 1\text{.}\) So by Definition 3.5.3, the series has radius of convergence \(1\text{.}\)
- The Taylor series representation of \(\log(1+x)\) is given in I. The series converges for \(-1\lt x\le 1\text{.}\) In particular, it converges for all \(|x|\lt 1\) and diverges for all \(|x|\gt 1\text{.}\) So by Definition 3.5.3, the series has radius of convergence \(1\text{.}\)
- The Taylor series representation of \(\arctan x\) is given in IV. The series converges for \(-1\le x\le 1\text{.}\) In particular, it converges for all \(|x|\lt 1\) and diverges for all \(|x|\gt 1\text{.}\) So it has radius of convergence \(1\text{.}\)
- The Taylor series representation of \(e^x\) is given in VI. The series converges for all \(x\text{.}\) So it has infinite radius of convergence.
- The Taylor series representation of \(\sin x\) is given in II. The series converges for all \(x\text{.}\) So it has infinite radius of convergence.
- The Taylor series representation of \(\cos x\) is given in III. The series converges for all \(x\text{.}\) So it has infinite radius of convergence.
3.6.8.4.
Solution.
- Using the definition of a Taylor series, we know\begin{equation*} \displaystyle\sum_{n=0}^\infty \frac{n^2}{(n!+1)}(x-3)^n=\displaystyle\sum_{n=0}^\infty \frac{f^{(n)}(3)}{n!}(x-3)^n \end{equation*}So, the coefficient of \((x-3)^{20}\) is \(\frac{f^{(20)}(3)}{20!}\) (using the definition). Using the given series, the coefficient of \((x-3)^{20}\) is \(\frac{20^2}{20!+1}\text{.}\) So,\begin{align*} \frac{f^{(20)}(3)}{20!}&=\frac{20^2}{20!+1}\\ \Rightarrow \qquad f^{(20)}(3)&=20^2\left(\frac{20!}{20!+1}\right) \end{align*}(which is extremely close to \(20^2\)).
- Using the definition of a Taylor series, we know\begin{equation*} \displaystyle\sum_{n=0}^\infty \frac{n^2}{(n!+1)}(x-3)^{2n}=\displaystyle\sum_{k=0}^\infty \frac{g^{(k)}(3)}{k!}(x-3)^{k} \end{equation*}So, the coefficient of \((x-3)^{20}\) is \(\frac{g^{(20)}(3)}{20!}\) (using the definition). Looking at the given series, the coefficient of \((x-3)^{20}\) occurs when \(n=10\text{,}\) so it is \(\frac{10^2}{10!+1}\text{.}\) So,\begin{align*} \frac{g^{(20)}(3)}{20!}&=\frac{10^2}{10!+1}\\ \Rightarrow \qquad g^{(20)}(3)&=10^2\left(\frac{20!}{10!+1}\right) \end{align*}
- With the previous two examples in mind, we find the Maclaurin series for \(h(x)\text{.}\) (Using the series representation will be much easier than differentiating \(h(x)\) directly twenty times.) Recall from the text that we know the Maclaurin series for \(\arctan x\text{.}\)\begin{align*} \arctan(x)&=\sum_{n=0}^\infty (-1)^n\frac{x^{2n+1}}{2n+1}\\ \arctan(5x^2)&=\sum_{n=0}^\infty (-1)^n\frac{{(5x^2)}^{2n+1}}{2n+1} =\sum_{n=0}^\infty (-1)^n\frac{5^{2n+1}}{2n+1}x^{4n+2}\\ \frac{\arctan(5x^2)}{x^2}&=\sum_{n=0}^\infty (-1)^n\frac{5^{2n+1}}{2n+1}x^{4n}\\ \sum_{k=0}^\infty \frac{h^{(k)}(0)}{k!}x^k&=\sum_{n=0}^\infty (-1)^n\frac{5^{2n+1}}{2n+1}x^{4n} \end{align*}Using the definition of a Maclaurin series, the coefficient of \(x^{20}\) is \(\dfrac{h^{(20)}(0)}{20!}\text{.}\) This occurs in the given series when \(n=5\text{,}\) so\begin{align*} \dfrac{h^{(20)}(0)}{20!}&=(-1)^5\frac{5^{2\times5+1}}{2\times5+1}=-\frac{5^{11}}{11}\\ \Rightarrow\qquad h^{(20)}(0)&=-\frac{20!\cdot 5^{11}}{11} \end{align*}Similarly, the coefficient of \(x^{22}\) in the Maclaurin series is \(\dfrac{h^{(22)(0)}}{22!}\text{.}\) Since no term \(x^{22}\) occurs in our series, that coefficient is 0, so \(h^{(22)}(0)=0\text{.}\)
3.6.8.5.
Solution.
3.6.8.6.
Solution.
3.6.8.7.
Solution.
3.6.8.8.
Solution.
3.6.8.9. (✳).
Solution.
3.6.8.10. (✳).
Solution.
3.6.8.11. (✳).
Solution.
3.6.8.12. (✳).
Solution.
3.6.8.13. (✳).
Solution.
3.6.8.14. (✳).
Solution.
3.6.8.15. (✳).
Solution.
3.6.8.16. (✳).
Solution.
3.6.8.17. (✳).
Solution.
3.6.8.18. (✳).
Solution.
3.6.8.19. (✳).
Solution.
3.6.8.20. (✳).
Solution.
3.6.8.21. (✳).
Solution.
3.6.8.22. (✳).
Solution.
3.6.8.23.
Solution.
3.6.8.24.
Solution.
3.6.8.25. (✳).
Solution.
- Solution 1: The naive strategy is to set \(a_n=\dfrac{x^{2n}}{(2n)!}\) and apply the ratio test.\begin{align*} \lim_{n\rightarrow\infty}\Big|\frac{a_{n+1}}{a_n}\Big| &=\lim_{n\rightarrow\infty} \left|\frac{ \frac{x^{2n+2}}{(2n+2)!} } {\frac{x^{2n}}{(2n)!}}\right| =\left| \frac{x^{2n+2}}{x^{2n}}\cdot \frac{(2n)!}{(2n+2)(2n+1)(2n)!} \right|\\ &=\lim_{n\rightarrow\infty} \frac{x^2}{(2n+2)(2n+1)}\\ &=0 \end{align*}This is smaller than \(1\) no matter what \(x\) is. So the series converges for all \(x\text{.}\)
- Solution 2: Alternatively, the sneaky way is to observe that both \(e^x=\displaystyle \sum_{n=0}^\infty \frac{x^n}{n!}\) and \(e^{-x}=\displaystyle \sum_{n=0}^\infty \frac{(-x)^n}{n!}\) are known to converge for all \(x\text{.}\) So\begin{equation*} \frac{1}{2}\big(e^x+e^{-x}\big) = \sum_{{n{\rm\ even}}}\frac{x^n}{n!} = \sum_{n=0}^\infty \frac{x^{2n}}{(2n)!} \end{equation*}also converges for all \(x\text{.}\)
3.6.8.26.
Solution.
- Using the Taylor series for arctangent when \(x=1\text{,}\) we see\begin{align*} \frac{\pi}{4}=\arctan(1)&=\sum_{n=0}^\infty(-1)^n \frac{1}{2n+1}\\ \pi &=\sum_{n=0}^\infty (-1)^n\frac{4}{2n+1} \end{align*}The error involved in approximating \(\pi\) with the partial sum \(S_N\) is at most \(|a_{N+1}|=\frac{4}{2N+3}\text{.}\) In order for this to be at most \(4\times 10^{-5}\text{,}\) we need:\begin{align*} \frac{4}{2N+3}&\le 4\times 10^{-5}\\ 2N+3 &\ge 10^5\\ N&\ge \frac{10^5-3}{2}=5\times 10^4-\frac{3}{2}=50,000-1.5 \end{align*}Since \(n\) must be an integer, we need to add up the terms from \(n=0\) to \(n=49,999\text{.}\) That is, we add up the first 50,000 terms.
-
Using the Taylor series for arctangent:\begin{align*} \pi&=16\arctan\frac15-4\arctan\frac{1}{239}\\ \amp=16\sum_{n=0}^\infty(-1)^n\frac{1}{(2n+1)5^{2n+1}}- 4\sum_{n=0}^\infty(-1)^n\frac{1}{(2n+1)\cdot 239^{2n+1}}\\ &=\sum_{n=0}^\infty\frac{(-1)^n}{2n+1}\left(\frac{16}{5^{2n+1}}-\frac{4}{239^{2n+1}} \right) \end{align*}This is an alternating sum, so the absolute error in using the partial sum \(S_N\) is at most:\begin{equation*} |a_{N+1}|=\frac{1}{2N+3}\left(\frac{16}{5^{2N+3}} -\frac{4}{239^{2N+3}}\right) \end{equation*}So, we want to find a value of \(N\) that makes this at most \(4\times 10^{-5}\text{.}\) Several values of \(N\) are given below.
\(N\) \(|a_{N+1}|\) 1 \(\dfrac{1}{5}\left(\dfrac{16}{5^5}-\dfrac{4}{239^5} \right)\approx 0.001\) 2 \(\dfrac{1}{7}\left(\dfrac{16}{5^7}-\dfrac{4}{239^7} \right)\approx 0.000029 \lt 4\times 10^{-5}\) So, it suffices to add up the first three terms (\(n=0\text{,}\) \(n=1\text{,}\) and \(n=2\)) of the series. -
Again, we use the Taylor series for arctangent.\begin{align*} \arctan\frac12+\arctan\frac13&=\arctan\left(\frac{3+2}{2\cdot3-1}\right)=\arctan(1)=\frac{\pi}{4} \end{align*}so that\begin{align*} \pi&=4\left(\arctan\frac12+\arctan\frac13 \right)\\ &=4\sum_{n=0}^\infty (-1)^n\frac{1}{(2n+1)2^{2n+1}}+4\sum_{n=0}^\infty (-1)^n\frac{1}{(2n+1)3^{2n+1}}\\ &=\sum_{n=0}^\infty (-1)^n\frac{4}{2n+1}\left(\frac{1}{2^{2n+1}}+\frac{1}{3^{2n+1}} \right) \end{align*}If we use the partial sum \(S_N\text{,}\) our absolute error is at most\begin{equation*} |a_{N+1}|=\dfrac{4}{2N+3}\left(\dfrac{1}{2^{2N+3}}+\dfrac{1}{3^{2N+3}} \right). \end{equation*}Several of these values are given below.
\(N\) \(|a_{N+1}|\) 1 \(\displaystyle \frac{4}{5}\left(\frac{1}{2^5}+\frac{1}{3^5} \right)\approx 0.028\) 2 \(\displaystyle\frac{4}{7}\left(\frac{1}{2^7}+\frac{1}{3^7} \right)\approx 0.0047\) 3 \(\displaystyle\frac{4}{9}\left(\frac{1}{2^9}+\frac{1}{3^9} \right)\approx 0.00089\) 4 \(\displaystyle\frac{4}{11}\left(\frac{1}{2^{11}}+\frac{1}{3^{11}} \right)\approx 0.00018\) 5 \(\displaystyle\frac{4}{13}\left(\frac{1}{2^{13}}+\frac{1}{3^{13}} \right)\approx 0.000038 \lt 4\times10^{-5}\) So, it suffices to add the first six terms (\(n=0\) to \(n=5\)) of the series.
3.6.8.27.
Solution.
\(N\) | \(|a_{N+1}|\) |
10 | \(\displaystyle \frac{1}{11\cdot 2^{11}}\approx 4\times 10^{-5}\) |
15 | \(\displaystyle\frac{1}{16\cdot 2^{16}}\approx 9.5\times 10^{-7}\) |
20 | \(\displaystyle\frac{1}{21\cdot 2^{21}}\approx 2 \times 10^{-8}\) |
25 | \(\displaystyle\frac{1}{26\cdot 2^{26}}\approx 6 \times 10^{-10}\) |
26 | \(\displaystyle\frac{1}{27\cdot 2^{27}}\approx 3 \times 10^{-10}\) |
27 | \(\displaystyle\frac{1}{28\cdot 2^{28}}\approx 1 \times 10^{-10}\) |
28 | \(\displaystyle\frac{1}{29\cdot 2^{29}}\approx 6 \times 10^{-11}\) |
29 | \(\displaystyle\frac{1}{30\cdot 2^{30}}\approx 3 \times 10^{-11}\) |
3.6.8.28.
Solution.
\(N\) | \(\dfrac{3}{(N+1)!}\) |
10 | \(\displaystyle \frac{3}{11!}=\frac{1}{9^{10}}\approx 8\times 10^{-8}\) |
11 | \(\displaystyle \frac{3}{12!}\approx 6\times 10^{-9}\) |
12 | \(\displaystyle \frac{3}{13!}\approx 5\times 10^{-10}\) |
13 | \(\displaystyle \frac{3}{14!}\approx 3\times 10^{-11}\) |
3.6.8.29.
Solution.
\(N\) | \(\dfrac{1}{9^{N+1}\cdot (N+1)}\) |
8 | \(\displaystyle \frac{1}{9\cdot 9^{9}}=\frac{1}{9^{10}}\approx 3\times 10^{-10}\) |
9 | \(\displaystyle \frac{1}{10\cdot 9^{10}}\approx 3\times 10^{-11}\) |
3.6.8.30.
Solution.
\(N\) | \(\dfrac{9\cdot 2^{N+1}}{(N+1)!}\) |
10 | \(\displaystyle \frac{9\cdot 2^{11}}{(11)!}\approx 5\times 10^{-4}\) |
15 | \(\displaystyle \frac{9\cdot 2^{16}}{(16)!}\approx 3\times 10^{-8}\) |
17 | \(\displaystyle \frac{9\cdot 2^{18}}{(18)!}\approx 4\times 10^{-10}\) |
18 | \(\displaystyle \frac{9\cdot 2^{19}}{(19)!}\approx 4\times 10^{-11} \lt 5\times10^{-11}\) |
3.6.8.31.
Solution.
- \(\displaystyle E(0)=\dfrac{-5^7}{14\cdot 6^7}[2]\)
- \(\displaystyle E\left(-\frac{1}{3}\right)=\dfrac{-5^7}{14\cdot 6^7}\left[\left(\frac43 \right)^{-7}+\left(\frac23 \right)^{-7} \right]=\dfrac{-5^7}{14\cdot 6^7}\left[\left(\frac34 \right)^{7}+\left(\frac32 \right)^{7} \right]\)
- \(\displaystyle E\left(\frac{1}{2}\right)=\dfrac{-5^7}{14\cdot 6^7}\left[\left(\frac12 \right)^{-7}+\left(\frac32 \right)^{-7} \right]=\dfrac{-5^7}{14\cdot 6^7}\left[2^7+\left(\frac23 \right)^{7} \right]\)
3.6.8.32. (✳).
Solution.
3.6.8.33. (✳).
Solution.
3.6.8.34.
Solution.
3.6.8.35.
Solution.
3.6.8.36.
Solution.
3.6.8.37.
Solution.
We can find the antiderivative of arctangent using integration by parts. Let \(u=\arctan x\) and \(\dee{v}=\dee{x}\text{;}\) then \(\dee{u}=\frac{1}{1+x^2}\dee{x}\) and \(v=x\text{.}\)
\begin{align*} \int \arctan x \,\dee{x}&=x\arctan x - \int \frac{x}{1+x^2}\,\dee{x}+C\\ \end{align*}Now, we use the substitution \(w=1+x^2\text{,}\) \(\dee{w}=2x\dee{x}\text{.}\)
\begin{align*} &=x\arctan x - \frac{1}{2}\log(1+x^2)+C\\ \text{So, } \sum_{n=0}^\infty (-1)^n\frac{x^{2n+2}}{(2n+1)(2n+2)}&=x\arctan x - \frac{1}{2}\log(1+x^2)+C\\ \end{align*}To find \(C\text{,}\) we evaluate both sides of the equation at \(x=0\text{.}\)
\begin{align*} 0&=0\arctan 0 -\frac{1}{2}\log(1)+C=C\\ \text{Therefore, } \sum_{n=0}^\infty (-1)^n\frac{x^{2n+2}}{(2n+1)(2n+2)}&=x\arctan x - \frac{1}{2}\log(1+x^2)\\ \end{align*}Multiplying both sides by \(x^2\text{,}\)
\begin{align*} \sum_{n=0}^\infty (-1)^n\frac{x^{2n+4}}{(2n+1)(2n+2)}&=x^3\arctan x - \frac{x^2}{2}\log(1+x^2) \end{align*}3.6.8.38.
Solution.
- We’ll start, as we usually do, by finding a pattern for \(f^{(n)}(0)\text{.}\)\begin{align*} f(x)&=(1-x)^{-1/2}\\ f'(x)&=\frac{1}{2}(1-x)^{-3/2}\\ f''(x)&=\frac{1\cdot3}{2^2}(1-x)^{-5/2}\\ f'''(x)&=\frac{1\cdot3\cdot5}{2^3}(1-x)^{-7/2}\\ f^{(4)}(x)&=\frac{1\cdot3\cdot5\cdot7}{2^4}(1-x)^{-9/2}\\ \vdots&\\ f^{(n)}(x)&=\frac{1\cdot3\cdot5\cdot \ldots \cdot (2n-1)}{2^n}(1-x)^{-(2n+1)/2}\\ f^{(n)}(0)&=\frac{1\cdot3\cdot5\cdot \ldots \cdot (2n-1)}{2^n}\\ \end{align*}This pattern holds for \(n\ge0\text{.}\) Now, we can write our Maclaurin series for \(f(x)\text{.}\)
We could leave it like this, but we simplify, to make our work cleaner later on.
\begin{align*} &=\frac{1}{2^n}\cdot\frac{(2n)!}{2\cdot4\cdot6\cdot\ldots\cdot(2n)}\\ &=\frac{1}{2^n}\cdot\frac{(2n)!}{2^n\cdot n!}\\ &=\frac{(2n)!}{2^{2n}\,n!} \end{align*}\begin{equation*} (1-x)^{-1/2}=\sum_{n=0}^\infty\frac{f^{(n)}(0)}{n!}x^n=\sum_{n=0}^\infty\frac{(2n)!}{2^{2n}\,(n!)^2}x^n \end{equation*}To find the radius of convergence, we use the ratio test.\begin{align*} \left|\frac{a_{n+1}}{a_n}\right| &=\frac{(2n+2)!}{2^{2n+2}((n+1)!)^2}\cdot\frac{2^{2n}\,(n!)^2}{(2n)!}\cdot|x|\\ &=\frac{(2n+2)!}{(2n)!}\left(\frac{n!}{(n+1)!} \right)^2\cdot\frac{2^{2n}}{2^{2n+2}}|x|\\ &=(2n+2)(2n+1)\left(\frac{1}{n+1} \right)^2\cdot\frac{1}{4}|x|\\ &=\frac{4n^2+4n+2}{4n^2+8n+4}|x|\\ \lim_{n \to \infty}\left|\frac{a_{n+1}}{a_n}\right|&=\lim_{n \to \infty}\left[\frac{4n^2+4n+2}{4n^2+8n+4}|x| \right]=|x| \end{align*}So, the radius of convergence is \(R=1\text{.}\) - We note the derivative of the arcsine function is \(\dfrac{1}{\sqrt{1-x^2}}=f(x^2)\text{.}\) With this insight, we can manipulate our Taylor series for \(f(x)\) into a Taylor series for arcsine.\begin{align*} \frac{1}{\sqrt{1-x}}&=\sum_{n=0}^\infty\frac{(2n)!}{2^{2n}\,(n!)^2}x^n\\ \frac{1}{\sqrt{1-x^2}}&=\sum_{n=0}^\infty\frac{(2n)!}{2^{2n}\,(n!)^2}x^{2n}\\ \int\frac{1}{\sqrt{1-x^2}}\,\dee{x}&=\int\left(\sum_{n=0}^\infty\frac{(2n)!}{2^{2n}\,(n!)^2}x^{2n}\right)\,\dee{x}\\ \arcsin x &=\sum_{n=0}^\infty\frac{(2n)!}{2^{2n}\,(n!)^2(2n+1)}x^{2n+1}+C\\ \arcsin x &=\sum_{n=0}^\infty\frac{(2n)!}{2^{2n}\,(n!)^2(2n+1)}x^{2n+1} \end{align*}where we found the value of \(C\) by setting \(x=0\text{.}\) Its radius of convergence is also 1, by Theorem 3.5.13.
3.6.8.39. (✳).
Solution.
3.6.8.40. (✳).
Solution.
3.6.8.41. (✳).
Solution.
3.6.8.42. (✳).
Solution.
\(n\) | 0 | 1 | 2 | 3 |
\(\dfrac{(-1)^n}{n!(2n+3)}\frac{1}{2^{2n+3}}\) | 0.04167 | -0.00625 | 0.00056 | -0.00004 |
3.6.8.43. (✳).
Solution.
3.6.8.44. (✳).
Solution.
3.6.8.45. (✳).
Solution.
3.6.8.46. (✳).
Solution.
3.6.8.47. (✳).
Solution.
3.6.8.48. (✳).
Solution.
3.6.8.49. (✳).
Solution.
3.6.8.50.
Solution.
- For Newton’s method, recall we approximate a root of the function \(g(x)\) in iterations: given an approximation \(x_n\text{,}\) our next approximation is \(x_{n+1}=x_n-\dfrac{g(x_n)}{g'(x_n)}\text{.}\) In our case,\begin{equation*} x_{n+1}=x_n - \frac{x_n^3-2}{3x_n^2}=\frac23\left(x_n+\frac{1}{x_n^2}\right). \end{equation*}We want to start somewhere reasonably close to the actual root we want, so let’s set \(x_0=1\text{.}\) (Your starting point may vary.)\begin{align*} x_0&=1 &\implies x_1\amp=\dfrac23\left(1+\frac11\right)=\dfrac{4}{3}\\ &&\amp\approx 1.3333\\ x_1&=\dfrac43 &\implies x_2\amp=\dfrac23\left(\frac43+\frac{9}{16}\right) =\dfrac{91}{72}\\ &&\amp\approx1.2639\\ x_2&=\dfrac{91}{72} &\implies x_3\amp=\dfrac23\left(\dfrac{91}{72}+\dfrac{72^2}{91^2}\right)=\dfrac{1126819 }{894348 }\\ &&\amp\approx1.2599\\ x_3&=\dfrac{1126819 }{894348 } &\implies x_4\amp=\dfrac23\left(\dfrac{1126819 }{894348 }+\dfrac{894348^2}{1126819^2}\right)\\ &&\amp\approx1.2599 \end{align*}So, \(\sqrt[3]{2}\approx 1.26\text{.}\)
-
We’ll evaluate the given series at \(x=2\text{.}\) This yields the series\begin{equation*} \sqrt[3]{2}=1+\frac{1}{6}+\sum_{n=2}^\infty(-1)^{n-1}\frac{(2)(5)(8)\cdots(3n-4)}{3^n\, n!}. \end{equation*}This series is alternating, so if we use the partial sum \(S_N\text{,}\) our absolute error is at most\begin{equation*} |a_{N+1}|=\frac{(2)(5)(8)\cdots(3N-1)}{3^{N+1}\, (N+1)!} \end{equation*}(if \(N \ge 2\)). We want to know which value of \(N\) makes this at most 0.01. We test several values.
\(N\) \(|a_{N+1}|\) 3 \(\dfrac{(2)(5)(8)}{3^4\cdot 4!}\approx 0.04\) 4 \(\dfrac{(2)(5)(8)(11)}{3^5\, 5!}\approx 0.03\) 5 \(\dfrac{(2)(5)(8)(11)(14)}{3^6\, 6!}\approx 0.023\) 6 \(\dfrac{(2)(5)(8)(11)(14)(17)}{3^7\, 7!}\approx 0.019\) 7 \(\dfrac{(2)(5)(8)(11)(14)(17)(20)}{3^8\, 8!}\approx 0.016\) 8 \(\dfrac{(2)(5)(8)(11)(14)(17)(20)(23)}{3^9\, 9!}\approx 0.013\) 9 \(\dfrac{(2)(5)(8)(11)(14)(17)(20)(23)(26)}{3^{10}\, 10!}\approx 0.012\) 10 \(\dfrac{(2)(5)(8)(11)(14)(17)(20)(23)(26)(29)}{3^{11}\, 11!}\approx 0.0103\) 11 \(\dfrac{(2)(5)(8)(11)(14)(17)(20)(23)(26)(29)(32)}{3^{12}\, 12!}\approx 0.009\) So, the approximation \(S_{11}\) has a sufficiently small error. That is, we would add up the first twelve terms.
3.6.8.51.
Solution.
- Make a Taylor series for \(f(x)\)
- Calculate the tenth derivative of the Taylor series of \(f(x)\text{.}\)
- Decide how many terms we need to add to achieve the desired accuracy.
- Approximate \(f^{(10)}\left(\frac15\right)\) with a partial sum.
3.6.8.52.
Solution.
-
To sketch \(y=f(x)\text{,}\) we note the following:
- \(f(x)\) is never negative.
- \(\lim\limits_{x \to \pm \infty} f(x)=e^0=1\text{,}\) so the curve has horizontal asymptotes in both directions at \(y=1\text{.}\)
- \(\lim\limits_{x \to \pm 0} f(x)=\lim\limits_{x \to \pm 0} \frac{1}{e^{1/x^2}}=\lim\limits_{u \to +\infty}\frac{1}{e^u}=0=f(0)\text{,}\) so the curve is continuous at \(x=0\text{.}\)
- For \(x\neq 0\text{,}\) \(f'(x)=\frac{2}{x^3}e^{-1/x^2}\text{,}\) so our curve is decreasing on \((-\infty,0)\) and increasing on \((0,\infty)\)
- For \(x\neq 0\text{,}\) \(f''(x)=2x^{-6}(2-3x^2)e^{-1/x^2}\text{,}\) so our curve is concave up on \((-\sqrt{2/3},\sqrt{2/3})\text{,}\) and concave down elsewhere.
- Since \(f^{(n)}(0)=0\) for all whole \(n\) (that is, the graph is really quite flat at the origin), and since \(f(0)=0\text{,}\) the Maclaurin series for \(f(x)\) is \(\displaystyle\sum_{n=0}^\infty \frac{0}{n!}x^n=0\text{.}\)
- The Maclaurin series converges for all real values of \(x\) (to the constant 0).
- Since \(e^y \gt 0\) for any real \(y\text{,}\) we see \(f(x)=0\) only when \(x=0\text{.}\) So, \(f(x)\) is only equal to its Maclaurin series at the single point \(x=0\text{.}\)
3.6.8.53.
Solution.
- Solution 1: Since \(f(x)\) is odd, \(f(-x)=-f(x)\) for all \(x\) in its domain. We plug this into our power series, then consider the even-indexed terms and the odd-indexed terms separately.\begin{align*} f(-x)&=-f(x)\\ \sum_{n=0}^\infty\frac{f^{(n)}(0)}{n!}(-x)^n&=-\sum_{n=0}^\infty\frac{f^{(n)}(0)}{n!}x^n \end{align*}Now separating the even powered and odd powered terms\begin{align*} \amp\textcolor{red}{\sum_{n=0}^\infty\frac{f^{(2n+1)}(0)}{(2n+1)!}(-x)^{2n+1}}+ \textcolor{blue}{\sum_{n=0}^\infty\frac{f^{(2n)}(0)}{(2n)!}(-x)^{2n}}\\ &\hskip0.5in =\textcolor{red}{-\sum_{n=0}^\infty\frac{f^{(2n+1)}(0)}{(2n+1)!}x^{2n+1}}\textcolor{blue}{-\sum_{n=0}^\infty\frac{f^{(2n)}(0)}{(2n)!}x^{2n}} \end{align*}For any integer \(n\text{,}\) we have that \((-1)^{2n}=1\) and \((-1)^{2n+1}=-1\) so that\begin{align*} \amp\textcolor{red}{-\sum_{n=0}^\infty\frac{f^{(2n+1)}(0)}{(2n+1)!}x^{2n+1}}+ \textcolor{blue}{\sum_{n=0}^\infty\frac{f^{(2n)}(0)}{(2n)!}x^{2n}}\\ &\hskip0.5in=\textcolor{red}{-\sum_{n=0}^\infty\frac{f^{(2n+1)}(0)}{(2n+1)!}x^{2n+1}}\textcolor{blue}{-\sum_{n=0}^\infty\frac{f^{(2n)}(0)}{(2n)!}x^{2n}} \end{align*}and\begin{align*} \textcolor{blue}{\sum_{n=0}^\infty\frac{f^{(2n)}(0)}{(2n)!}x^{2n}}&=\textcolor{blue}{-\sum_{n=0}^\infty\frac{f^{(2n)}(0)}{(2n)!}x^{2n}} \end{align*}and\begin{align*} 2\textcolor{blue}{\sum_{n=0}^\infty\frac{f^{(2n)}(0)}{(2n)!}x^{2n}}&=0\\ \textcolor{blue}{\sum_{n=0}^\infty\frac{f^{(2n)}(0)}{(2n)!}x^{2n}}&=0 \end{align*}
- Solution 2: Alternately, we could note the following:
- Since all derivative of \(f(x)\) exist, all its derivatives are continuous.
- The derivative of an odd function is even, and the derivative of an even function is odd.
- So, the even-indexed derivatives of \(f(x)\) are continuous, odd functions.
- Every continuous, odd function passes through the origin. That is, \(f^{(2n)}(0)=0\text{.}\)
- So, every term in the series is \(0\text{.}\)