Example 34:
Evaluate
Using ,
Adding (i) and (ii),
Using ,
Put . Then .
When , ; when , .
Using when , and ,
Putting this in (iii),
.
Want the reason behind every step, the basics and the checks?
Nobody can write down an antiderivative of with the functions of this course, so the fundamental theorem is no help. Instead we make the integral reproduce itself. P4 says the integral of equals that of ; adding them gives ; and the integral of turns out (by a substitution and P6) to be the original integral again. That leaves an equation , which we solve for .
- Given
- , where is the natural logarithm (base ), as everywhere in this chapter.
- To find
- The exact value of .
- Plan
- (1) Deal with the endpoint , where is undefined, and show is a finite number. (2) P4: . (3) Add: is the integral of . (4) Substitute . (5) P6 on gives back . (6) Solve.
Basics you need
P4: . P6: if . Lesson: properties of definite integrals
For : , and . Only positive numbers have logarithms. Basics: logarithms
, , and . Basics: trigonometric identities
Change the variable, write the new differential, and change the limits to the values of the new variable. Lesson: integration by substitution
Solution
The trouble at , and why is still a finite number
On we have , so is defined, continuous and . But at , has no logarithm, and as , . So the integrand is not continuous on the closed interval , and Theorem 2 of Section 7.8.3 does not apply directly (Remark (iv) there asks for a function continuous on ). Such an integral is called improper: it means the limit of as . The textbook does not raise this point, but the method below needs to be a finite number, so we check it.
The graph of is concave on , so it lies above its chord from to :
Since is increasing, for :
With (so ) and :
using as . So the negative function is squeezed between , whose integral is finite, and 0. Its integral therefore converges, and
This matters in Step 6, where we subtract from both sides of an equation. That is only allowed for a finite number: "" tells us nothing. The same bound (reflected or rescaled) shows that every other integral below, of , and on , is also finite. The properties P0 to P7 and the substitution rule then hold for them too, because each one can be applied on a slightly shorter closed interval, where everything is continuous, before letting the cut-off shrink to 0.Concave means on : the curve bends downwards, so it stays above any chord. The antiderivative of is found by parts in Section 7.6 (check: its derivative is ).P4: the same integral with in place of
P4 with upper limit replaces by :
Replacing by reads the interval backwards, so the area does not change. On this interval it turns into : the graph of is the mirror image of the graph of in the line (see the figure in Step 6).Add the two forms of
Now the book adds and subtracts and asks "Why?". The reason is the double-angle formula: . So
Integrating each term (the constant integrates to over an interval of length ):
For both and are positive, so is a correct use of the product law. We want a single sine, because a single sine can be turned back into the original integral; the factor is exactly what is missing to become .Substitute in the first integral
Put . Then , so . When , ; when , . Hence
and
The substitution makes the argument of a single variable again, at the price of doubling the interval to . The map is increasing and takes exactly onto , so it is a legitimate change of variable on the whole interval. On , , so is defined there.Section 7.9: change the limits along with the variable, then there is no need to substitute back. Basics: substitutionP6 folds back onto
Take and , so . The condition of P6 holds:
So P6 gives
where the last equality is just P0: the name of the variable does not matter. Therefore
The graph of on is symmetric about , so the two halves and contribute equally. The left half is our original integral. This is the moment the integral "reproduces itself".(supplementary angles). P0: .Solve for
Subtract from both sides (allowed because is a finite number, Step 1):
This lies between and 0, as Step 1 promised. It is negative because on the whole interval: the region lies below the -axis.
(blue) and its mirror image (dashed violet) on . The shaded region between the blue curve and the axis lies below the axis, so it counts as negative area: . The curve plunges to at (it is drawn down to ), yet the shaded area is finite. The final answer can also be read as: the average value of over is , the same as .
(an improper but convergent integral). By Step 2 the same value holds for .
Numerical integration (adaptive quadrature, which copes with the logarithmic endpoint) gives , and .
It also confirms the intermediate results: (Step 5) and (Steps 4 and 5 together). At the midpoint, .
- Splitting into . The log of a product is the sum of the logs.
- Forgetting the factor from , or forgetting to change the upper limit from to . Either slip destroys the cancellation in Step 5.
- Applying P6 without checking . Here it holds because ; with on it would fail (, and is not even defined for ).
- Ignoring the endpoint. Solving assumes is finite. For an integral that diverges, the same algebra would produce a meaningless "answer".
- When no antiderivative is available, look for a chain of properties that brings the integral back to itself, then solve the resulting equation for .
- P4 () and P6 (folding in half) are the two tools; a substitution like links an integral on with one on .
- Remember the value: It is used again in Exercise 7.10, Q10 and Q16. It also gives at once, since on .
Every fact used, and where it comes from
| Fact | Used in | Source |
|---|---|---|
| A definite integral needs continuous on | Step 1 | Remark (iv) after Theorem 2, Section 7.8.3 |
| on | Step 1 | Concavity of on (a concave curve lies above its chords); beyond this textbook |
| Step 1 | Integration by parts, Section 7.6 | |
| P4: | Step 2 | Section 7.10, property P4 (Lesson: properties) |
| ; | Step 3 | Laws of logarithms; Class XI trigonometry |
| Substitution with changed limits | Step 4 | Section 7.9 |
| P6 with ; P0 | Step 5 | Section 7.10, properties P6 and P0 |
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 →