Example 21:
Find
Let .
Taking as the first function, by parts:
Again by parts, with as the first function:
Using (ii) in (i):
.
Want the reason behind every step, the basics and the checks?
Neither factor ever disappears: stays , and cycles through , , . So integrating by parts once gives , and integrating by parts again brings back , the integral we started with. Calling it , we get an equation for , and we simply solve it.
- Given
- , defined and continuous for every real .
- To find
- .
- Plan
- (1) By parts once ( first, second), giving . (2) By parts on with the same kind of choice. (3) Substitute and solve for . (4) See why the choice must be kept. (5) Differentiate back.
Basics you need
(Section 7.6). Lesson: integration by parts
; , (standard formulae (iii) and (ii)). Basics: trigonometric functions
Two antiderivatives of the same function differ by a constant (Section 7.2), so an equation between indefinite integrals holds "up to a constant"; the constant is added at the end. Basics: integration undoes differentiation
Solution
Integrate by parts once
Take as the first function and as the second. Then the derivative of the first is and the integral of the second is :
Write , so
Here neither choice makes the integral simpler; it only turns into . That is fine: the point is to do it twice and come back to .(formula (iii)). Lesson: integration by partsIntegrate by parts, keeping as the first function
Now is first and second; the integral of is :
The last integral is exactly the original . That is not a failure; it is what makes the method work, because now appears on both sides of an equation.(formula (ii)). Lesson: integration by partsSubstitute and solve for
Put into (1):
Add to both sides:
We treat like an unknown in an ordinary linear equation. Strictly, each stands for an antiderivative, determined only up to a constant, so the equation holds up to a constant; that is why one arbitrary constant is added after solving.Why the choice must be kept the second time
Suppose in Step 2 we had switched, taking first and second. Then
and (1) would become , that is : true, but useless.
Switching the roles in the second round simply undoes the first round. Keep the same type of function (here the exponential) as the first function both times. The book also notes that taking first and second in both rounds works equally well and gives the same answer.Integration by parts is the product rule read backwards; applying it forwards and then backwards returns to the start. Lesson: integration by partsCheck: differentiate the answer
By the product rule: differentiates to itself, and differentiates to . The cosines cancel and the sines add, giving exactly the integrand for every real .
valid for every real . This is the book's answer.
At a point. At : the integrand is ; the derivative from Step 5 is , the same value.
Over an interval. On :
Numerical integration of over gives to ten decimal places, the same.
- Switching the first and second functions in the second round. You get (Step 4).
- Thinking the method has failed when reappears. Move it to the left side and solve.
- Sign slips. , so the first round gives , with a plus in front of the integral.
- Forgetting to divide by 2 after collecting .
- For or : integrate by parts twice with the same choice, then solve the equation for .
- ; similarly .
Every fact used, and where it comes from
| Fact | Used in | Source |
|---|---|---|
| Integration by parts | Steps 1, 2, 4 | Section 7.6 (Lesson: integration by parts) |
| ; | Steps 1, 2 | Standard formulae (iii), (ii), Section 7.2 |
| Antiderivatives differ by a constant | Step 3 | Section 7.2 |
| Product rule; derivatives of , , | Step 5 | Chapter 5 |
Tell us if something on this page is wrong or unclear, or if you would like a fuller explanation, another method or a new feature. Send feedback on this problem →