8

Techniques of Integration

Over the next few sections we examine some techniques that are frequently successful when
seeking antiderivatives of functions. Sometimes this is a simple problem, since it will be
apparent that the function you wish to integrate is a derivative in some straightforward
way. For example, faced with
x10 dx
we realize immediately that the derivative of x11 will supply an x10 : (x11 )′ = 11x10 . We
don’t want the “11”, but constants are easy to alter, because differentiation “ignores” them
in certain circumstances, so
d 1 11
1
x =
11x10 = x10 .
dx 11
11
From our knowledge of derivatives, we can immediately write down a number of antiderivatives. Here is a list of those most often used:

xn dx =

xn+1
+ C,
n+1

if n = −1

x−1 dx = ln |x| + C
ex dx = ex + C
sin x dx = − cos x + C
163

164

Chapter 8 Techniques of Integration

cos x dx = sin x + C
sec2 x dx = tan x + C
sec x tan x dx = sec x + C
1
dx = arctan x + C
1 + x2
1

dx = arcsin x + C
1 − x2

º½

ËÙ ×Ø ØÙØ ÓÒ

Needless to say, most problems we encounter will not be so simple. Here’s a slightly more
complicated example: find
2x cos(x2 ) dx.
This is not a “simple” derivative, but a little thought reveals that it must have come from
an application of the chain rule. Multiplied on the “outside” is 2x, which is the derivative
of the “inside” function x2 . Checking:
d
d
sin(x2 ) = cos(x2 ) x2 = 2x cos(x2 ),
dx
dx
so
2x cos(x2 ) dx = sin(x2 ) + C.
Even when the chain rule has “produced” a certain derivative, it is not always easy to
see. Consider this problem:
x3

1 − x2 dx.

There are two factors in this expression, x3 and 1 − x2 , but it is not apparent that the
chain rule is involved. Some clever rearrangement reveals that it is:
x3

1 − x2 dx =

(−2x) −

1
2

(1 − (1 − x2 )) 1 − x2 dx.

This looks messy, but we do now have something that looks like the result of the chain

rule: the function 1 − x2 has been substituted into −(1/2)(1 − x) x, and the derivative

1 Substitution 165 of 1 − x2 . multiplied on the outside. u. then do some algebra to see what this requires the rest of the function to look like. So finally we have x3 1 − x2 dx = 1 1 (1 − x2 ) − 5 3 (1 − x2 )3/2 + C. u −2 dx Recall that one benefit of the Leibniz notation is that it often turns out that what looks like ordinary arithmetic gives the correct answer. It does sometimes not work. but the idea is simple: guess at the most likely candidate for the “inside function”. One frequently good guess is any complicated expression inside a square root. Now we know that the chain rule will multiply by the derivative of this inner function: du = −2x. without requiring cleverness to rewrite a function in just the right way.1. for convenience in the manipulations that follow. or may require more than one attempt. dx so we need to rewrite the original function to include this: x3 1 − x2 = √ −2x x3 u dx = −2x x2 √ du dx.1) +C x3/2 + C. rewriting the original function so that it looked like the result of using the chain rule. So we succeeded. since then 1 d F (1 − x2 ) = −2xF ′ (1 − x2 ) = (−2x) − dx 2 = x3 (1 − (1 − x2 )) 1 − x2 1 − x2 But this isn’t hard: √ 1 − (1 − x) x dx = 2 1 − (x1/2 − x3/2 ) dx 2 2 3/2 2 5/2 x − x 3 5 =− 1 2 = 1 1 x− 5 3 (8. there is a technique that makes such problems simpler. even if something more complicated is . so we start by trying u = 1 − x2 .8. using a new variable. If we can find a function F (x) whose derivative √ is −(1/2)(1 − x) x we’ll be done. Fortunately. but it required a clever first step. −2x.

If we can integrate this new function of u. we have simply renamed the variable u to make the calculations less confusing. for example. then translate the given function so that it is written entirely in terms of u. and . For example. we let u denote a likely candidate for the inner function. To summarize: if we suspect that a given function is the derivative of another via the chain rule. then the antiderivative of the original function is obtained by replacing u by the equivalent expression in x.166 Chapter 8 Techniques of Integration going on. dx dt dx The same is true of our current expression: x2 √ du u dx = −2 dx x2 √ u du. is: Since du/dx = 2x. consider again this simple problem: 2x cos(x2 ) dx. For example. we can replace it by du: 2x cos(x2 ) dx = cos u du = sin u + C = sin(x2 ) + C. −2 Now we’re almost there: since u = 1 − x2 . with no x remaining in the expression. This is not the only way to do the algebra.1. x2 = 1 − u and the integral is √ 1 − (1 − u) u du. Even in simple cases you may prefer to use this mechanical procedure.1). then du/dx = 2x or du = 2x dx. Let u = x2 . since it often helps to avoid silly mistakes. Another possibility. dx = du/2x. and typically there are many paths to the correct answer. in Leibniz notation the chain rule is dy dt dy = . 2 It’s no coincidence that this is exactly the integral we computed in (8. Then since u = 1 − x2 : x3 1 − x2 dx = 1 1 (1 − x2 ) − 5 3 (1 − x2 )3/2 + C. Since we have exactly 2x dx in the original integral. Just as before: √ 1 − (1 − u) u du = 2 1 1 u− 5 3 u3/2 + C.

a = 0. It is dangerous. u = 16. (ax + b)n dx.1. and dangerous. Let u = x2 so du = 2x dx or x dx = du/2. then Evaluate EXAMPLE 8.2 sin(ax + b) dx. The important thing to remember is that you must eliminate all instances of the original variable x. 2 2 1 sin u du means that u takes on values between 2 and 4. assuming that a and b are constants.1. Then 1 1 1 sin u du = (− cos u) + C = − cos(ax + b) + C. assuming that a and b are constants and a = 0. u = 4. Since u = x2 .8. alternative is something like this: 4 4 2 x sin(x ) dx = 2 2 1 1 sin u du = − cos(u) 2 2 4 This is incorrect because 2 4 2 1 = − cos(x2 ) 2 4 2 1 1 = − cos(16) + cos(4). Then 1 1 1 sin u du = (− cos u) + C = − cos(x2 ) + C. when x = 2. We let u = ax + b so du = a dx or dx = du/a. Evaluate EXAMPLE 8. Then (ax + b)n dx = 1 n 1 1 u du = un+1 + C = (ax + b)n+1 + C.1. which 2 1 is wrong.1 and n is a positive integer. because it is very easy to get to the point − cos(u) 2 4 and forget 2 . 2 2 An incorrect. a a(n + 1) a(n + 1) Evaluate EXAMPLE 8.3 2 evaluate the definite integral. Again we let u = ax + b so du = a dx or dx = du/a. 2 2 2 x sin(x2 ) dx = Now 4 1 x sin(x ) dx = − cos(x2 ) 2 4 2 2 2 1 1 = − cos(16) + cos(4). First we compute the antiderivative. So we can do this: 4 16 x sin(x2 ) dx = 2 4 1 1 sin u du = − (cos u) 2 2 16 4 1 1 = − cos(16) + cos(4). 2 2 A somewhat neater alternative to this method is to change the original limits to match the variable u. a a a sin(ax + b) dx = 4 x sin(x2 ) dx. and when x = 4.1 Substitution 167 then the integral becomes 2x cos(x2 ) dx = 2x cos u du = 2x cos u du.

sec2 x tan x dx ⇒ 14. tan x dx ⇒ 12. x(x2 + 1)100 dx ⇒ 4. sin3 x cos x dx ⇒ 6. Find the antiderivatives or evaluate the definite integral in each problem. sin5 (3x) cos(3x) dx ⇒ π/2 x sec2 (x2 ) tan(x2 ) dx ⇒ 1 dx ⇒ (3x − 7)2 6x dx ⇒ (x2 − 7)1/9 −1 sin7 x dx ⇒ 1 dt ⇒ 1 − 5t 100 − x2 dx ⇒ 10.1. Then EXAMPLE 8. π 11. thus getting the incorrect answer − cos(4) + cos(2). √ 3 5.168 Chapter 8 Techniques of Integration 1 1 to substitute x2 back in for u. 1. A 2 2 somewhat clumsy.4 1/2 1/4 Evaluate cos(πt) dt = sin2 (πt) 1 √ 2/2 1 1 du = π u2 1 √ 2/2 1 −2 1 u−1 u du = π π −1 1 √ 2/2 √ 2 1 . 0 (cos2 x − sin2 x) dx ⇒ 1 18. (1 − t)9 dt ⇒ 2.1. (x2 + 1)2 dx ⇒ 20. −1 1 19. 2 2 1/2 cos(πt) dt. 0 4 15. Let u = sin(πt) so du = π cos(πt) dt or 2 1/4 sin (πt) √ du/π = cos(πt) dt. =− + π π Exercises 8. (2x3 − 1)(x4 − 2x)6 dx ⇒ f (x)f ′ (x) dx ⇒ . sin(tan x) dx ⇒ cos2 x π/6 16. but acceptable. 9. alternative is something like this: 4 x=4 x sin(x2 ) dx = 2 x=2 x=4 1 1 sin u du = − cos(u) 2 2 x=2 4 1 = − cos(x2 ) 2 2 =− cos(16) cos(4) + . cos(πt) cos sin(πt) dt ⇒ x2 √ dx ⇒ 1 − x3 sin x dx ⇒ cos3 x 7. 3. 3 17. We change the limits to sin(π/4) = 2/2 and sin(π/2) = 1. x 8. 0 √ 13.

Now we have four integrals to evaluate: 1 dx = x and 3 −3 cos 2x dx = − sin 2x 2 .8. Use sin2 x = (1 − cos(2x))/2 to rewrite the function: sin6 x dx = (1 − cos 2x)3 dx 8 (sin2 x)3 dx = = 1 8 1 − 3 cos 2x + 3 cos2 2x − cos3 2x dx.2. Some examples will suffice to explain the approach.2 Evaluate sin6 x dx. These can sometimes be tedious. Now use u = cos x. 3 5 EXAMPLE 8. but the technique is straightforward.2 º¾ ÈÓÛ Ö× Ó × Ò Ò Powers of sine and cosine 169 Ó× Ò Functions consisting of products of the sine and cosine can be integrated by using substitution and trigonometric identities.1 Evaluate sin5 x dx = sin5 x dx. Rewrite the function: sin x sin4 x dx = sin x(sin2 x)2 dx = sin x(1 − cos2 x)2 dx. du = − sin x dx: sin x(1 − cos2 x)2 dx = = −(1 − u2 )2 du −(1 − 2u2 + u4 ) du 2 1 = −u + u3 − u5 + C 3 5 1 2 = − cos x + cos3 x − cos5 x + C. EXAMPLE 8.2.

170 Chapter 8 Techniques of Integration are easy. cos2 x sin3 x dx ⇒ 5.3 Evaluate sin 2x − sin3 2x 3 + 3 16 x+ sin 4x 4 + C. sin4 x dx ⇒ 4. sec2 x csc2 x dx ⇒ 10. Exercises 8. Use the formulas sin2 x = (1−cos(2x))/2 and cos2 x = (1 + cos(2x))/2 to get: sin2 x cos2 x dx = 1 − cos(2x) 1 + cos(2x) · dx. The cos3 2x integral is like the previous example: − cos3 2x dx = − cos 2x cos2 2x dx = − cos 2x(1 − sin2 2x) dx = 1 − (1 − u2 ) du 2 =− 1 2 u− =− 1 2 sin 2x − u3 3 sin3 2x 3 . cos3 x dx ⇒ 6. Find the antiderivatives.2. 2 2 The remainder is left as an exercise. sin2 x cos2 x dx. sin2 x dx ⇒ 2. And finally we use another trigonometric identity.2. sin x(cos x)3/2 dx ⇒ 9. cos3 x sin2 x dx ⇒ 8. cos2 x = (1 + cos(2x))/2: 3 cos2 2x dx = 3 1 + cos 4x 3 dx = 2 2 x+ sin 4x 4 . sin3 x dx ⇒ 3. 1. So at long last we get sin6 x dx = x 3 1 − sin 2x − 8 16 16 EXAMPLE 8. sin2 x cos2 x dx ⇒ 7. tan3 x sec x dx ⇒ .

8. try x = sin u. since √ cos2 u is positive. try x = sec u. This seems like a “reverse” substitution. but it is really no different in principle than ordinary substitution. though the term sin(2 arcsin x) is a bit unpleasant. Then 1 − sin2 u cos u du = √ cos2 u cos u du. Using the identity sin 2x = 2 sin x cos x. we can write sin 2u = 2 sin u cos u = 2 sin(arcsin x) full antiderivative is 1 − sin2 u = 2x 1 − sin2 (arcsin x) = 2x 1 − x2 .3 º¿ ÌÖ Trigonometric Substitutions 171 ÓÒÓÑ ØÖ ËÙ ×Ø ØÙØ ÓÒ× So far we have seen that it sometimes helps to replace a subexpression of a function by a single variable. Consider again the substitution x = sin u. We could just as well think of this as u = arcsin x. −π/2 ≤ u ≤ π/2. so cos u ≥ 0. if it contains 1 + x2 try x = tan u. If your function contains 1 − x2 . but this is valid only if cos u is positive. as in the example above.1 Evaluate 1 − x2 dx = 1 − x2 dx. Sometimes you will need to try something a bit different to handle constants other than one. Then we continue: √ cos2 u cos u du = = cos2 u du = u sin 2u 1 + cos 2u du = + +C 2 2 4 arcsin x sin(2 arcsin x) + + C.3. 2 4 2 2 This type of substitution is usually indicated when the function you wish to integrate contains a polynomial expression that might allow you to use the fundamental identity sin2 x + cos2 x = 1 in one of three forms: cos2 x = 1 − sin2 x sec2 x = 1 + tan2 x tan2 x = sec2 x − 1. 2 4 This is a perfectly good answer. and if it contains x2 − 1. EXAMPLE 8. Then the √ √ arcsin x x 1 − x2 arcsin x 2x 1 − x2 + = + + C. Occasionally it can help to replace the original variable by something more complicated. If we do. Let x = sin u so dx = cos u du. . It is possible to simplify this. √ We would like to replace cos2 u by cos u. then by the definition of the arcsine.

1.3.3. −π/2 ≤ u ≤ π/2 and sec u ≥ 0. Then 2 1 − (3x/2)2 dx = 2 1 − sin2 u (2/3) cos u du = 4 3 cos2 u du 4u 4 sin 2u + +C 6 12 2 arcsin(3x/2) 2 sin u cos u = + +C 3 3 2 arcsin(3x/2) 2 sin(arcsin(3x/2)) cos(arcsin(3x/2)) = + +C 3 3 = 2 arcsin(3x/2) 2(3x/2) 1 − (3x/2)2 + +C 3 3 √ 2 arcsin(3x/2) x 4 − 9x2 = + + C.3 Evaluate 1 + x2 dx. Then sec3 u du.3. two integrals come up frequently: sec3 u du and Both have relatively nice expressions but they are a bit tricky to discover. 3 2 = using some of the work from example 8. so √ sec2 u sec2 u du = √ √ sec2 u sec2 u du. EXAMPLE 8. sec2 u = sec u.172 Chapter 8 Techniques of Integration EXAMPLE 8.2 Evaluate 4 − 9x2 dx. sec u du. Let x = tan u. We start by rewriting this so that it looks more like the previous example: 4 − 9x2 dx = 4(1 − (3x/2)2 ) dx = 1 − (3x/2)2 dx. . 2 Now let 3x/2 = sin u so (3/2) dx = cos u du or dx = (2/3) cos u du. dx = sec2 u du. so 1 + tan2 u sec2 u du = 1 + x2 dx = Since u = arctan(x). In problems of this type.

so we just need the first quotient. sec3 u du: Now for sec3 u sec3 u sec3 u (tan2 u + 1) sec u sec u = + = + 2 2 2 2 3 2 3 sec u sec u tan u sec u sec u + sec u tan2 u sec u = + + = + . = 2 2 = using tan(arctan x) = x and sec(arctan x) = 1 + tan2 (arctan x) = 1 + x2 . Thus sec u du = sec2 u + sec u tan u du = sec u + tan u 1 dw = ln |w| + C w = ln | sec u + tan u| + C. 2 2 and reverting to the original variable x: 1 + x2 dx = sec u tan u ln | sec u + tan u| + +C 2 2 sec(arctan x) tan(arctan x) ln | sec(arctan x) + tan(arctan x)| + +C 2 2 √ √ ln | 1 + x2 + x| x 1 + x2 + + C. which we will need to compute sec u du = = sec u 173 sec3 u du: sec u + tan u du sec u + tan u sec2 u + sec u tan u du.8. . exactly the numerator of the function we are integrating. sec u + tan u Now let w = sec u + tan u. 2 2 2 2 2 3 We already know how to integrate sec u.3 First we do Trigonometric Substitutions sec u du. So putting these together we get sec3 u du = sec u tan u ln | sec u + tan u| + + C. This is “simply” a matter of recognizing the product rule in action: sec3 u + sec u tan2 u du = sec u tan u. dw = sec u tan u + sec2 u du.

1 √ dx ⇒ 1 + x2 1 dx ⇒ 2 (1 + x2 ) x √ x √ dx ⇒ 1−x ÁÒØ csc3 x dx ⇒ x2 8. when we noticed that sec3 u + sec u tan2 u du = sec u tan u. and is usually written in more compact form. Start with the product rule: d f (x)g(x) = f ′ (x)g(x) + f (x)g ′ (x). 5. Ö Ø ÓÒ 1 − x2 dx ⇒ x2 + 2x dx ⇒ x2 √ dx ⇒ 4 − x2 x3 √ dx ⇒ 4x2 − 1 Ý È ÖØ× We have already seen that recognizing the product rule can be useful. but it turns out that in many cases we have an integral of the form f (x)g ′ (x) dx but that f ′ (x)g(x) dx is easier. csc x dx ⇒ x2 − 1 dx ⇒ 3.3. 7. 10. If we let u = f (x) and v = g(x) then . there is a technique that will often help to uncover the product rule.174 Chapter 8 Techniques of Integration Exercises 8. 9. 12. f (x)g ′ (x) dx = f (x)g(x) − f ′ (x)g(x) dx. Find the antiderivatives. 1. This technique for turning one integral into another is called integration by parts. x 9 + 4x2 dx ⇒ 4. 1 − x2 dx ⇒ 6. 11. f (x)g(x) = and then This may not seem particularly useful at first glance. As with substitution. º 2. dx We can rewrite this as f ′ (x)g(x) dx + f (x)g ′ (x) dx. we do not have to rely on insight or cleverness to discover such antiderivatives.

EXAMPLE 8. but we needed to be clever to discover it. Then we must let dv = x dx so v = x2 /2 and x ln x dx = EXAMPLE 8.4. Then du = sec x tan x dx and v = tan x and sec3 x dx = sec x tan x − tan2 x sec x dx = sec x tan x − (sec2 x − 1) sec x dx = sec x tan x − sec3 x dx + sec x dx.1 Evaluate x ln x dx. To use this technique we need to identify likely candidates for u = f (x) and dv = g ′ (x) dx. Then we must let dv = sin x dx so v = − cos x and x sin x dx = −x cos x − EXAMPLE 8. Let u = x so du = dx.3 − cos x dx = −x cos x + Evaluate cos x dx = −x cos x + sin x + C. sec3 x dx.4. Here we’ll use the new technique to discover the antiderivative. 2 2 4 x sin x dx. .2 x2 ln x − 2 x2 1 x2 ln x dx = − 2 x 2 Evaluate x x2 ln x x2 dx = − + C.4. Of course we already know the answer to this.4 Integration by Parts 175 du = f ′ (x) dx and dv = g ′ (x) dx and u dv = uv − v du. Let u = ln x so du = 1/x dx. Let u = sec x and dv = sec2 x dx.8.

but it can be a bit tedious to accomplish. Let u = 2x. dv = sin x dx. Here is the table: sign u dv u dv x2 sin x x2 sin x − 2x − cos x −2x − cos x − 0 0 cos x 2 − sin x cos x or 2 − sin x . then du = 2 and v = sin x. Such repeated use of integration by parts is fairly common.4 sec x tan x 1 + 2 2 sec x dx 2x cos x dx. But looking more closely: At first this looks useless—we’re right back to sec3 x dx = sec x tan x − sec3 x dx + sec3 x dx + sec3 x dx = sec x tan x + sec x dx 2 sec3 x dx = sec x tan x + sec x dx sec3 x dx = = sec x tan x ln | sec x + tan x| + + C. dv = cos x dx. Let u = x2 . Now sec x dx x2 sin x dx. and x2 sin x dx = −x2 cos x + 2x cos x dx = −x2 cos x + 2x sin x − 2 sin x dx = −x2 cos x + 2x sin x + 2 cos x + C. We illustrate with the previous example. but we need to do integration by parts again.176 Chapter 8 Techniques of Integration sec3 x dx.4. There is a nice tabular method to accomplish the calculation that minimizes the chance for error and speeds up the whole process. This is better than the original integral. then du = 2x dx Evaluate EXAMPLE 8. 2 2 x2 sin x dx = −x2 cos x + and v = − cos x. and it is easy to make errors. especially sign errors involving the subtraction in the formula.

To compute with this second table we begin at the top. and combine these as −x2 cos x + 2x sin x − 2 sin x dx. (2)(− sin x). as above. if you do this by hand. we start with u at the top of the second column and repeatedly compute the derivative. Now we multiply twice on the diagonal. and (2)(cos x). we place a “−” in every second row. Typically we would fill in the table one line at a time. starting with dv at the top of the third column. computing the products as indicated will then give the entire integral. x2 cos x dx ⇒ 3. we may return yet once more to the table. We combine these as before to get −x2 cos x + 2x sin x + 2 cos x + 0 dx = −x2 cos x + 2x sin x + 2 cos x + C. While this integral is easy. (x2 )(− cos x) and (−2x)(− sin x) and then once straight across. If we can see that the u column will eventually become zero. xex dx ⇒ 4. sin2 x dx ⇒ 6.4 Integration by Parts 177 To form the first table. xex dx ⇒ 5. (0)(cos x). ln x dx ⇒ 2 . 1. Find the antiderivatives. In the first column. Now multiply three times on the diagonal to get (x2 )(− cos x). and once straight across.8. Since this integral is not yet easy. Exercises 8. Multiply the first entry in column u by the second entry in column dv to get −x2 cos x. To form the second table we combine the first and second columns by ignoring the boundary. or exactly the result of the first application of integration by parts.4. you may simply start with two columns and add a “−” to every second row. we repeatedly compute the antiderivative. This gives: −x2 cos x + 2x cos x dx. (−2x)(− sin x). we can instead fill in the whole table. including the “+C ”. we return to the table. giving the same result as the second application of integration by parts. x cos x dx ⇒ 2. until the “straight across” multiplication gives an easy integral. and add this to the integral of the product of the second entry in column u and second entry in column dv.

178 Chapter 8 Techniques of Integration 8. √ arctan( x) dx ⇒ 13. So we shall explain how to find the antiderivative of a rational function only when the denominator is a quadratic polynomial ax2 + bx + c. something that is not always possible. allows us to integrate any rational function. . x3 sin x dx ⇒ x3 cos x dx ⇒ 10. However. in principle. The denominator becomes un . While it may be tedious to complete the integration if the numerator has high degree. x sin x cos x dx ⇒ 12. x2 + x − 6 (x − 3)2 x2 − 1 are all rational functions of x. 1 x2 + 1 x3 . in practice one does not often run across rational functions with high degree polynomials in the denominator for which one has to find the antiderivative function. x arctan x dx ⇒ 9. x sin2 x dx ⇒ 11. The algebraic steps in the technique are rather cumbersome if the polynomial in the denominator has degree more than 2. sec2 x csc2 x dx ⇒ 7. º Ê Ø ÓÒ Ð ÙÒ Ø ÓÒ× A rational function is a fraction with polynomials in the numerator and denominator. √ sin( x) dx ⇒ 14. and dx = du/a. We should mention a special type of rational function that we already know how to integrate: If the denominator has the form (ax + b)n . it is merely a matter of algebra. and the technique requires that we factor the denominator. . For example. and each x in the numerator is replaced by (u − b)/a. . There is a general technique called “partial fractions” that. the substitution u = ax + b will always work.

Since there is no square root of −3.8.2 Determine whether x2 + x + 1 factors. depending on how the quadratic factors: either x2 + bx + c = (x − r)(x − s).5. EXAMPLE 8. so we can assume that the denominator has the form x2 + bx + c. 2 2 √ √ 1+ 5 x− 2 √ 1− 5 x− 2 . EXAMPLE 8. this quadratic does not factor. and factor it if possible. and to factor the quadratic if it is possible. We can always factor out the coefficient of x2 and put it outside the integral. Using the substitution u = 3 − 2x we get (3 − 2x)5 Find EXAMPLE 8. The quadratic formula tells us that x2 − x − 1 = 0 when x= Therefore x2 − x − 1 = 1± √ 1+4 1± 5 = .5.5 1 x3 dx = (3 − 2x)5 −2 = 1 16 = 1 16 = 1 16 =− x3 dx.5. We can use the quadratic formula to decide which of these we have. The quadratic formula tells us that x2 + x + 1 = 0 when x= −1 ± √ 2 1−4 . or it doesn’t factor. x2 + bx + c = (x − r)2 . and factor it if possible. .1 179 Rational Functions u−3 −2 u5 3 du = 1 16 u3 − 9u2 + 27u − 27 du u5 u−2 − 9u−3 + 27u−4 − 27u−5 du u−1 9u−2 27u−3 27u−4 − + − −1 −2 −3 −4 +C (3 − 2x)−1 9(3 − 2x)−2 27(3 − 2x)−3 27(3 − 2x)−4 − + − −1 −2 −3 −4 +C 1 9 9 27 + − + +C 2 3 16(3 − 2x) 32(3 − 2x) 16(3 − 2x) 64(3 − 2x)4 We now proceed to the case in which the denominator is a quadratic polynomial.3 Determine whether x2 − x − 1 factors. There are three possible cases.

An example should make it clear how to proceed. adding two fractions with constant numerator and denominators (x−r) and (x−s) produces a fraction with denominator (x − r)(x − s) and a polynomial of degree less than 2 for the numerator. We want to end up with 7x − 6 A B = + . This is a problem you’ve seen before: solve a .5. The first step is to make sure that p(x) has degree less than 2. (x − 2)(x + 3) The first integral is easy.5 Evaluate dx.4 Rewrite so x3 x3 7x − 6 7x − 6 = 2 =x−1+ 2 =x−1+ . (x − 2)(x + 3) x−2 x+3 If we go ahead and add the fractions on the right hand side we get (A + B)x + 3A − 2B 7x − 6 = .5. The other two cases require different approaches. Now consider the following simple algebra of fractions: B A(x − s) + B(x − r) (A + B)x − As − Br A + = = . we need 7 = A + B and −6 = 3A − 2B. that can be handled with a substitution. (x − 2)(x + 3) x +x−6 x +x−6 (x − 2)(x + 3) x3 dx = (x − 2)(x + 3) x − 1 dx + 7x − 6 dx. x3 7x − 6 EXAMPLE 8. x3 dx in terms of an integral with a numerator (x − 2)(x + 3) that has degree less than 2. which is to say. so only the second requires some work. We want to reverse this process: starting with a single fraction. we have an integral of the form p(x) dx (x − r)(x − s) where p(x) is a polynomial. x−r x−s (x − r)(x − s) (x − r)(x − s) That is. we want to write it as a sum of two simpler fractions.180 Chapter 8 Techniques of Integration If x2 + bx + c = (x − r)2 then we have the special case we have already seen. (x − 2)(x + 3) (x − 2)(x + 3) So all we need to do is find A and B so that 7x − 6 = (A + B)x + 3A − 2B. To do this we use long division of polynomials to discover that EXAMPLE 8. We start by writing (x − 2)(x + 3) (x − 2)(x + 3) as the sum of two fractions. If x2 + bx + c = (x − r)(x − s).

2 1 dx = arctan +1 2 The final answer is now 1 1 x+1 dx = ln |x2 + 4x + 8| − arctan 2 + 4x + 8 x 2 2 x+2 2 x+2 2 . x+1 dx. The quadratic denominator does not + 4x + 8 factor. 5x−2 5 x+3 5 5 The answer to the original problem is now x3 dx = (x − 2)(x + 3) = x − 1 dx + 7x − 6 dx (x − 2)(x + 3) 8 27 x2 − x + ln |x − 2| + ln |x + 3| + C. Again we can use long division to ensure that the numerator has degree less than 2. This is easy to solve for A: A = 8/5. then we complete the square. 2 5 5 Now suppose that x2 + bx + c doesn’t factor. + C. making the integral 1 4 Using u = x+2 we get 2 1 1 1 dx = 2 x+2 4 4 +1 2 1 x+2 2 2 u2 +1 dx. We could complete the square and use a trigonometric substitution.8. 2 + 4x + 8 2 + 4x + 8 2 + 4x + 8 x x x EXAMPLE 8. .5 Rational Functions 181 system of two equations in two unknowns. There are many ways to proceed. Thus 7x − 6 dx = (x − 2)(x + 3) 27 1 8 27 8 1 + dx = ln |x − 2| + ln |x + 3| + C. and then B = 7 − A = 7 − 8/5 = 27/5.6 Evaluate x2 The first integral is an easy substitution problem. 2 + 4x + 8 x 2 u 2 For the second integral we complete the square: x+2 2 x2 + 4x + 8 = (x + 2)2 + 4 = 4 2 +1 .5. using u = x2 + 4x + 8: x+2 1 du 1 dx = = ln |x2 + 4x + 8|. here’s one: If 7 = A+B then B = 7−A and so −6 = 3A−2B = 3A−2(7−A) = 3A−14+2A = 5A−14. but it is simpler to rearrange the integrand: x+2 1 x+1 dx = dx − dx.

. ..182 Chapter 8 Techniques of Integration Exercises 8. . . ... dx ⇒ 4 + x2 x3 7.. .... we already know one way to approximate an integral: if we think of the integral as computing an area. . .. . . . .... . . ... While this is quite simple. .. 4 − x2 1 3. Unfortunately.. .. ... . . .. . .. .... . . . . . .... .. we divide the interval into n equal subintervals of length ∆x... ...... ... ......... .... .. ....... .. . . . . . . A f (xi ) + f (xi+1 ) typical trapezoid is pictured in figure 8... . 4. . ..6.6.. . . .. ........ . . . .1 we see an area under a curve approximated by rectangles and by trapezoids.. .. We will see two methods that work reasonably well and yet are fairly simple.. . ... . . ... .... . . . ...... . . . ..... .. .. . ... ........ . x4 dx ⇒ 4 − x2 x2 dx ⇒ 4 − x2 1 dx ⇒ 2 + 10x + 29 x 1 dx ⇒ x2 + 10x + 21 1 dx ⇒ 2 + 3x x Ö Ø ÓÒ We have now seen some of the most generally useful methods for discovering antiderivatives.... ... In figure 8. .. A similar approach is much better: we approximate the area under a curve over a small interval as the area of a trapezoid...... . ... . .... Figure 8. . . dx ⇒ x2 + 10x + 25 x4 5.. ... .. ... . . . ...... 6.. . . .... ... . . .. . ..... ..... Of course.. . it has area ∆x........ .. .. ... . ......6.. and there are others.. .. ..... .. . . . ... . . ... ... . . If we add up 2 . .. . . .. .. ... . . . .. it is usually the case that a large number of rectangles is needed to get acceptable accuracy. Find the antiderivatives.. . . . As with rectangles.. .. . . . . ... . .. . ... . .5. . ... . ... . .. . . .. . .... 1 dx ⇒ 1. . . .... . ... . ..... .. . . .. .. .. .. .... .. .1 . . .. . ... ...... . ... 8... .... .. . . . Approximating an area with rectangles and with trapezoids. . 10. . . . .. .. it is apparent that the trapezoids give a substantially better approximation on each subinterval. .. . .. ... . . . .. .. . ... .. . . .. ... ... .. . .. .. .. . ........ . ... . . . . . .... in such cases if the value of a definite integral is needed it will have to be approximated. .. . we can add up the areas of some rectangles... . ..... . in some cases more sophisticated techniques will be needed. .... ... .. . some functions have no simple antiderivatives. .. . .... . . ... . . . .. . .. . dx ⇒ 4 + x2 1 dx ⇒ 9. 2 −x−3 2x º ÆÙÑ Ö Ð ÁÒØ 2. ...2.

.8..1 Suppose f has a second derivative f ′′ everywhere on the interval [a. b].... .2 (xi+1 .. an approximation is useful only if we know how accurate it is. .. for many functions... In the case of our approximation of the integral.. .. and |f ′′ (x)| ≤ M for all x in the interval. f (xi+1 )) xi+1 A single trapezoid.. there is such an error estimate associated with the trapezoid approximation.. When we compute a particular approximation to an integral.. the error is the difference between the approximation and the true value of the integral.. we need an error estimate. we might need a particular value accurate to three decimal places. b−a (b − a)3 M (∆x)2 = M.. . . . If A is an approximation and E is the associated error estimate. 12 12n2 . . . .6 Numerical Integration 183 the areas of all trapezoids we get f (x1 ) + f (x2 ) f (xn−1 ) + f (xn ) f (x0 ) + f (x1 ) ∆x + ∆x + · · · + ∆x = 2 2 2 f (x0 ) f (xn ) ∆x..6.... . .. For any approximation technique. for example... . . then we know that the true value of the integral is between A − E and A + E.. xi Figure 8.. an error estimate for the trapezoid approximation is E(∆x) = Let’s see how we can use this... . .. . With ∆x = (b − a)/n.6.... . ... we want E = E(∆x) to be a function of ∆x that gets small rapidly as ∆x gets small. THEOREM 8. Fortunately. . . .. .. (xi .... . ... a computer can easily do many subintervals. In practice...... f (xi )) . ... ... . a value that is guaranteed to be larger than the actual error. . . .. For a modest number of subintervals this is not too difficult to do with a calculator. + f (x1 ) + f (x2 ) + · · · + f (xn−1 ) + 2 2 This is usually known as the Trapezoid Rule. ...

6.0047.77 ≈ 100 <n 3 With n = 6.001.2 2 e−x dx to two decimal places. The second deriva- Approximate 0 −x2 2 −x2 2 tive of f = e is (4x −2)e . We compute the trapezoid approximation for six intervals: f (0) f (1) + f (1/6) + f (2/6) + · · · + f (5/6) + 2 2 1 ≈ 0. especially compared to rectangles. We can extend this idea: what if we try to approximate the curve more closely. For comparison. . and it is not hard to see that on [0. using 12 rectangles to approximate the area gives 0. the first rounds to 0.0047 = 0. The trapezoid approximation works well.7727.005 12 n 1 (200) < n2 6 5. which turns out to be 0.001 12 n 1 (1000) < n2 6 12. Unfortunately. for example. because the tops of the trapezoids form a reasonably good approximation to the curve when ∆x is fairly small. In practice it generally pays to start by requiring better than the maximum possible error. the error estimate is thus 1/63 < 0.75. To get two decimal places of accuracy. we will certainly need E(∆x) < 0.74512. 6 So the true value of the integral is between 0. 1].005 or 1 1 (2) 2 < 0. which is considerably less accurate than the approximation using six trapezoids. we need to pick a larger n.74512 + 0.74 and the second rounds to 0. so we can’t be sure of the correct value in the second decimal place. We begin by estimating the number of subintervals we are likely to need.74042 and 0.0047 = 0. we might have initially required E(∆x) < 0. we need to go to n = 12 to get two bounds that both round to the same value.75.184 Chapter 8 Techniques of Integration 1 EXAMPLE 8.74512 − 0. |(4x2 −2)e−x | ≤ 2.91 ≈ 500 <n 3 Had we immediately tried n = 13 this would have given us the desired answer.74982. or 1 1 (2) 2 < 0. As it turns out.

and hope that the result is fairly simple. If we find a parabola through three consecutive points (xi . f (xi)). For this to be practical. we should attempt to write down the parabola y = ax2 + bx + c through these points and then integrate it. this turns out to be possible. 3 This is just slightly more complicated than the formula for trapezoids. and c: f (xi ) = a(xi+1 − ∆x)2 + b(xi+1 − ∆x) + c f (xi+1 ) = a(xi+1 )2 + b(xi+1 ) + c f (xi+2 ) = a(xi+1 + ∆x)2 + b(xi+1 + ∆x) + c Not surprisingly.8. Although the algebra involved is messy. and (xi+2 . (xi+1 . As with the trapezoid method. 3 Now the sum of the areas under all parabolas is ∆x (f (x0 ) + 4f (x1 ) + f (x2 ) + f (x2 ) + 4f (x3 ) + f (x4 ) + · · · + f (xn−2 ) + 4f (xn−1 ) + f (xn )) = 3 ∆x (f (x0 ) + 4f (x1 ) + 2f (x2 ) + 4f (x3 ) + 2f (x4 ) + · · · + 2f (xn−2 ) + 4f (xn−1 ) + f (xn )). Nevertheless. so we will present the result without all of the algebra. If we divide the interval [a. (xi+1 . xi+2 ]. we can then approximate the curve by a sequence of parabolas. (xi+2 . the parabola through (xi . The algebra is well within the capability of a good computer algebra system like Sage. this is useful only with an error estimate: . f (xi+2 )). f (xi+1 )).3. Sage can easily compute and simplify the integral to get xi+1 +∆x ax2 + bx + c dx = xi+1 −∆x ∆x (f (xi ) + 4f (xi+1 ) + f (xi+2 )). we need to remember the alternating 2 and 4 coefficients. f (xi+2 )) on the curve. each covering two of the subintervals. That is.6. as in figure 8. we can easily compute the area under the parabola. f (xi )).6 Numerical Integration 185 by using something other than a straight line? The obvious candidate is a parabola: if we can approximate a short piece of the curve with a parabola with equation y = ax2 + bx + c. b. it should be quite close to the curve over the whole interval [xi . f (xi+1 )). the solutions turn out to be quite messy. you can see how to do it in this Sage worksheet. namely. To find the parabola. This approximation technique is referred to as Simpson’s Rule. we solve these three equations for a. but only one through three given points. we would like a simple formula for the area under one parabola. There are an infinite number of parabolas through any two given points. b] into an even number of subintervals.

... .. To get two decimal places of accuracy. .. . .. With ∆x = (b − a)/n. ... .0003 = 0. 3·4 So the true value of the integral is between 0. . . ...... . 1] this is at most 12 in absolute value.. an error estimate for Simpson’s approximation is E(∆x) = b−a (b − a)5 M (∆x)4 = M. .... . . . . . .746855.. . . xi Figure 8.. .. .. .001 180 n 200 < n4 3 2.. both of which round to 0... since we need an even number of subintervals. . We begin by estimating the number of subintervals we are likely to need. .... ..746855 − 0.. and |f (4) (x)| ≤ M for all x in the interval. .... but taking a cue from our earlier example.0003 = 0. f (xi+2 )) xi+2 A parabola (dashed) approximating a curve (solid)...7471555.. THEOREM 8. ..6. ... .. .. f (xi )) ....... .. ... ..6..6..0003 and the approximation is (f (0) + 4f (1/4) + 2f (1/2) + 4f (3/4) + f (1)) 1 ≈ 0. . on [0.4 2 e−x dx to two decimal places.3 Suppose f has a fourth derivative f (4) everywhere on the interval [a... . . .. . .. .3 xi+1 (xi+2 . let’s require E(∆x) < 0... .. ..86 ≈ [4] 200 <n 3 So we try n = 4.. . .. ... .746555 and 0. . ..186 Chapter 8 Techniques of Integration (xi .. 180 180n4 1 EXAMPLE 8..... .. . . .75. . ...005... .... . . . . ..... .... ........ ..001: 1 1 (12) 4 < 0. . . .746855 + 0.... Then the error estimate is 12/180/44 < 0. The Let us again approximate 0 −x2 2 2 2 fourth derivative of f = e is (16x − 48x + 12)e−x .. .. b]. we will certainly need E(∆x) < 0... . . .

t(t2 − 9)3/2 dt ⇒ (et + 16)tet dt ⇒ 4. t sec2 t dt ⇒ 12. cos4 t dt ⇒ . 5. 1. 2 2 1 dt ⇒ − 4) cos 3t √ dt ⇒ sin 3t et √ dt ⇒ et + 1 t(t2 8. 2 2 5. Remarkably. you may use a graphing calculator or computer software to estimate the maximum values. 11. compute the trapezoid and Simpson approximations using 4 subintervals. In the following problems.6. 3 1 dx ⇒ x 1 1 √ x 1 + x dx ⇒ 0 1 x3 + 1 dx ⇒ 0 4 1 x4 + 1 dx ⇒ 9. and are in no particular order. You can use this to get started. because the parabolas used to approximate f will be f itself. f (x) dx = 3·2 x0 This does require a bit of messy algebra. Using Simpson’s rule on a parabola f (x). so you may prefer to use Sage. x3 dx ⇒ 1.) If you have access to Sage or similar software. 1 11. and compute the error estimate for each. Some problems may be done in more than one way. 3 3 x dx ⇒ 1 4. (t + 4)3 dt ⇒ 3. 7. 0 4 3. 1 x2 dx ⇒ 2. gives the exact value of the integral. 2t + 1 dt ⇒ +t+3 1 dt ⇒ (25 − t2 )3/2 t2 10. even with just two subintervals. 0 1 + 1/x dx ⇒ 10. º Ø ÓÒ Ð Ü Ö × × These problems require the techniques of this chapter.8. Simpson’s rule also computes the integral of a cubic function f (x) = ax3 + bx2 + cx + d exactly. x dx ⇒ 1+x 8. 1 5 7. 9. (Finding the maximum values of the second and fourth derivatives can be challenging for some of these. approximate each integral to two decimal places.7 Additional exercises 187 Exercises 8. sin t cos 2t dt ⇒ tan t sec2 t dt ⇒ 6. 1 dx ⇒ 1 + x2 6. 2. Show this is true by showing that x2 x2 − x0 (f (x0 ) + 4f ((x0 + x2 )/2) + f (x2 )).

t3 17. sin3 t cos4 t dt ⇒ 24. t2 + 1 dt ⇒ 1 dt ⇒ t(9 + 4t2 ) t dt ⇒ 2 + 2t − 3 t 1 dt ⇒ t2 − 6t + 9 t(ln t)2 dt ⇒ t2 t+1 dt ⇒ +t−1 . t2 √ 1 dt ⇒ 1 + t2 2 15. 23. 22. Chapter 8 Techniques of Integration t2 1 dt ⇒ + 3t 14. 25. sec t dt ⇒ (1 + tan t)3 16. √ (t3/2 + 47)3 t dt ⇒ 19. t3 et dt ⇒ 28. et sin t dt ⇒ 18. t3 dt ⇒ (2 − t2 )5/2 arctan 2t dt ⇒ 1 + 4t2 20. 1 dt ⇒ t(ln t)2 26. 21. 27.188 13.