Math 106: Review for Exam II - SOLUTIONS

INTEGRATION TIPS
• Substitution: usually let u = a function that’s “inside” another function, especially if du (possibly off
by a multiplying constant) is also present in the integrand.
Z Z Z Z
• Parts: u dv = uv − v du or uv0 dx = uv − u0v dx

How to choose which part is u? Let u be the part that is higher up in the LIATE mnemonic below.
(The mnemonics ILATE and LIPET will work equally well if you have learned one of those instead;
in the latter A is replaced by P, which stands for “polynomial”.)
Logarithms (such as ln x)
Inverse trig (such as arctan x, arcsin x)
Algebraic (such as x, x2, x3 + 4)
Trig (such as sin x, cos 2x)
Exponentials (such as ex , e3x)
• Rational Functions (one polynomial divided by another): if the degree of the numerator is greater than
or equal to the degree of the denominator, do long division then integrate the result.
Partial Fractions: here’s an illustrative example of the setup.

3x2 + 11 A B C Dx + E
2 2
= + + 2
+ 2
(x + 1)(x − 3) (x + 5) x + 1 x − 3 (x − 3) x +5

Each linear term in the denominator on the left gets a constant above it on the right; the squared
linear factor (x − 3) on the left appears twice on the right, once to the second power. Each irreducible
quadratic term on the left gets a linear term (Dx + E here) above it on the right.
• Trigonometric Substitutions: some suggested substitutions and useful formulae follow.
p p p
Radical Form a2 − x2 a2 + x2 x2 − a2
Substitution x = a sin t x = a tan t x = a sec t

sin2 x + cos2 x = 1 tan2 x + 1 = sec 2 x

1 cos(2x) 1 cos(2x)
sin2 x = − cos2 x = +
2 2 2 2
sin(2x) = 2 sin x cos x

• Powers of Trigonometric Functions: here are some strategies for dealing with these.

Z
sinm x cosn x dx Possible Strategy Identity to Use
m odd Break off one factor of sin x and substitute u = cos x. sin2 x = 1 − cos2 x
n odd Break off one factor of cos x and substitute u = sin x. cos2 x = 1 − sin2 x
1 cos(2x)
m, n even Use sin2 x + cos2 x = 1 to reduce to only powers of sin x sin2 x = −
2 2
1 cos(2x)
or only powers of cos x, then use table of integrals #39-42 cos2 x = +
2 2
or identities shown to right of this box.
 Z
tanm x secn x dx Possible Strategy Identity to Use
m odd Break off one factor of sec x tan x and substitute u = sec x. tan2 x = sec 2 x − 1
n even Break off one factor of sec 2 x and substitute u = tan x. sec2 x = tan2 x + 1
m even, n odd Use identity at right to reduce to powers of sec x alone. tan2 x = sec 2 x − 1
Then use table of integrals #51.

Useful Trigonometric Derivatives and Antiderivatives

Z
d d
tan x = sec2 x sec x = sec x tan x sec x dx = ln | sec x + tan x| + C
dx dx

• Improper integrals: look for ∞ as one of the limits of integration; look for functions that have a vertical
asymptote in the interval of integration. It may be useful to know the following limits.
lim ex = ∞
x→∞

lim e−x = 0 Note: this is the same as lim ex .

x→∞ x→−∞

lim 1/x = 0 Note: the answer is the same for lim 1/x2 and similar functions.
x→∞ x→∞

lim+ 1/x = ∞ Note: the answer is the same for lim+ 1/x2 and similar functions.
x→0 x→0
lim ln x = ∞
x→∞
lim ln x = −∞
x→0+
lim arctan x = π/2
x→∞

1. Evaluate the following.

(a) Let u = sin x, so du = cos x dx.
Z Z
sin6 x cos3 x dx = sin6 x(1 − sin2 x) cos x dx Use cos2 x = 1 − sin2 x.
Z
= u6(1 − u2) du
Z
= (u6 − u8 ) du

u7 u9
= − +C
7 9
sin7 x sin9 x
= − +C
7 9

(b) Let x = 10 tan t, so dx = 10 sec2 t dt.

p
 x2 + 102 = y2 ⇒ y = x2 + 100
y  √
 hyp x2 + 100
 x sec t = =
 adj 10
t opp x
10 tan t = =
adj 10
 Z Z
dx 10 sec2 t dt
√ = √
100 + x2 100 + 100 tan2 t
Z
10 sec2 t dt
= √ Now use 1 + tan2 t = sec2 t.
10 1 + tan2 t
Z
sec2 t dt
= √
Z sec2 t
= sec t dt

= ln | sec t + tan t| + C Now use triangle above.

√ 
 x2 + 100 x 

= ln  + +C
 10 10 

(c) This is an improper integral, so we need to use a limit.

Z ∞ Z t
1 1
100
dx = lim dx
3 x(ln x) t→∞ 3 x(ln x)100
Z x=t
1 dx
= lim du Substitute u = ln x, so du = .
t→∞ x=3 u100 x
u−99 x=t
= lim 
t→∞ −99 x=3
−1 t

= lim 
t→∞ 99(ln x)99 3
 
−1 −1
= lim −
t→∞ 99(ln t)99 99(ln 3)99
−1
= 0−
99(ln 3)99
1
= So, the integral converges (to this value).
99(ln 3)99

e−2x
(d) We’ll use integration by parts: u = x ⇒ du = dx and dv = e−2x ⇒ v = .
−2
Z ∞ Z t
−2x
xe dx = lim xe−2x dx
0 t→∞ 0
 −2x  Z t −2x 
e t e
= lim x  − dx
t→∞ −2 0 0 −2
 −2x t
e e−2x
= lim x −
t→∞ −2 4 0
 t
−x 1
= lim −
t→∞ 2e2x 4e2x 0
   
−t 1 0 1
= lim − − −
t→∞ 2e2t 4e2t 2e0 4e0
=(0 − 0) − (0 − 1/4)
=1/4 So, the integral converges (to this value).

(e) Partial Fractions:

3x2 + 2x − 13 Ax + B C
Write = 2 + . Now multiply both sides by (x − 3)(x2 + 1) to get
(x − 3)(x2 + 1) x +1 x−3
 3x2 + 2x − 13 = (Ax + B)(x − 3) + C(x2 + 1).
Let x = 3. Then 20 = C(10), so C = 2.
Let x = 0. Then −13 = B(−3) + 2(1), so B = 5.
Let x = 1. Then −8 = (A(1) + 5)(−2) + 2(2), so A = 1.

Z Z  
3x2 + 2x − 13 x+5 2
dx = + dx
(x − 3)(x2 + 1) x2 + 1 x − 3
Z  
x 5 2
= + + dx Let u = x2 + 1, so du = 2xdx.
x2 + 1 x2 + 1 x − 3
Z 1 Z  
2 du 5 2
= + + dx
u x2 + 1 x − 3
ln u
= + 5 arctan x + 2 ln |x − 3| + D
2
ln(x2 + 1)
= + 5 arctan x + 2 ln |x − 3| + D
2

(f) Since the degree of the numerator is greater than or equal to the degree of the denominator, we
do long division.
−5
4x2 − 3x + 2 +
x−6

x − 6 4x3 − 27x2 + 20x − 17
4x3 − 24x2
−3x2
−3x2 +18x
2x
2x − 12
−5
Now, we compute the integral.
Z Z  
4x3 − 27x2 + 20x − 17 2 5 4x3 3x2
dx = 4x − 3x + 2 − = − + 2x − 5 ln |x − 6| + C
x−6 x−6 3 2
(g) This integral is improper at x = 1 because the integrand has a vertical asymptote there.
Z 3 Z 3
1 1
dx = lim dx
1 x−1 t→1+ t x − 1
3
= lim ln |x − 1|t
t→1+
= lim [ln |3 − 1| − ln |t − 1|]
t→1+

Since lim (− ln |t − 1|) = ∞, this integral diverges (to ∞).

t→1+
2. Solve the differential equation dy/dx = 2xy + 6x if the solution passes through (0, 5)
5).
dy
= 2xy + 6x
dx
dy
= 2x(y + 3)
dx
dy
= 2x dx Separate the variables.
y +3
Z Z
dy
= 2x dx
y +3
ln |y + 3| = x2 + C
2
|y + 3| = ex +C
Exponentiate each side to remove the ln.
C x2
y + 3 = ±e e |w| = z means w = ±z.
2
y = −3 + Aex Replace ±eC with A.

Now we use the initial condition y(0) = 5 to find the value of A.

2
We have 5 = −3 + Ae0 ⇒ A = 8, so the solution is y = −3 + 8ex .
√
3. Find the second-degree Taylor polynomial for f(x) = x centered at x = 100
100.
f(x) = x1/2 f(100) = 10
1 1 1 1
f 0 (x) = x−1/2 = √ f 0 (100) = √ =
2 2 x 2 100 20
−1 −3/2 −1 −1 −1
f 00 (x) = x = 3/2 f 00 (100) = 3/2
=
4 4x 4 · 100 4000

f 00 (100)
P2(x) = f(100) + f 0 (100)(x − 100) + (x − 100)2
2!
x − 100 (x − 100)2
= 10 + −
20 8000

4. What is the maximum possible error that can occur in your Taylor approximation from
the previous problem on the interval [100, 110]?
Kn+1
We know that |f(x) − Pn(x)| ≤ |x − x0|n+1.
(n + 1)!
In this case, n = 2, x0 = 100, and x = 110 (the farthest from x0 that we are considering).
3 3 3
K3 = max of |f 000(x)| on [100, 110] = max of | 5/2 | on [100, 110] = =
8x 8 · 1005/2 800, 000
3
800,000 1
Putting this all together, we have |f(x) − P2 (x)| ≤ |110 − 100|3 = .
3! 1600
5. Use comparisons to show whether each of the following converges or diverges. If an
integral converges, also give a good upper bound for its value.
Z ∞
6 + cos x
(a) dx
1 x0.99
6 + cos x 6−1 5
For all x ≥ 1, we have 0.99
≥ 0.99 = 0.99 because the minimum value of cos x is −1.
Z ∞ x x x
5 dx
Since 0.99
diverges (compute yourself or notice that p = 0.99 < 1), we know that the integral
1 x
in question must diverge too.
 Z ∞
4x3 − 2x2
(b) dx
1 2x4 + x5 + 1
4x3 − 2x2 4x3 1
For all x ≥ 1, we have 4 5
≤ 5
= 4 2 . (We’ve made the denominator smaller and
2x + x + 1 x x
the numerator larger, so the new fraction is larger.)

Z ∞ Z t
dx dx
4 = 4 lim
1 x2 t→∞ 1 x
2

−1 t
= 4 lim 
t→∞ x 1
 
−1 −1
= 4 lim −
t→∞ t 1
= 4[0 − (−1)]
=4
Therefore, the original integral in question must converge to a value less than 4.

6. (Sections A and B may omit this question.) The probability density function (pdf ) of the
length (in minutes) of phone calls on a certain wireless network is given by f(x) = ke−0.2x
where x is the number of minutes. Note that the domain is x ≥ 0 since we can’t have a
negative number of minutes.

(a) What must be the value of k ?

We know that the total area under any pdf must be 1 (because it must account for 100% of
events.)
Z ∞ Z t
ke−0.2x dx = lim ke−0.2x dx
0 t→∞ 0
ke−0.2x t
= lim 
t→∞ −0.2 0
−0.2t
ke ke0
= lim −
t→∞ −0.2 −0.2
k
= 0−
−0.2
= 5k

So, we have 5k = 1 or k = 0.2.

(b) What fraction of calls last more than 3 minutes?
Z ∞ Z t
0.2e−0.2x dx = lim 0.2e−0.2x dx
3 t→∞ 3
0.2e−0.2x t
= lim 
t→∞ −0.2 3
= lim −e−0.2t − (−e−0.6 )
t→∞
= 0 + e−0.6
= e−0.6 ≈ 0.5488

Z 3
Note that we could instead have computed 1 − 0.2e−0.2x dx and gotten the same answer, but
0
the point of introducing pdf’s in this text seems to be to show how improper integrals are used.