Math 212 - Review for Exam 3

1. Answer each of the following questions as either True or False. Circle the correct answer.
∞
X
(a) True/False: If an > 0 and an ↓ 0, the series an converges.
n=1
∞
X
Soln: False: Let an = 1/n. Clearly an → 0 but the series an is the harmonic series which, by the integral test, is divergent.
n=1

∞
X ∞
X
(b) True/False: If an converges, then an converges.
n=10 n=1
Soln: True: One series is just a constant plus the other one. If one converges, so does the other.
∞
X
(c) True/False: If an > 0 for n = 1, 3, 5, · · · and an < 0 for n = 2, 4, 6, 8, · · · , then an converges.
n=1
∞
Soln: False: an = (−1)n+1 satisfies the above alternating property but the series
X
an is equal to 1 − 1 + 1 − 1 + 1 − 1 · · · which does not
n=1
converge since the terms do not go to zero.
∞
X sin x n
(d) True/False: The series ( ) converges for all x.
n=1
2
sin x
Soln: True: The above series is a geometric series with r = . Since −1 ≤ sin x ≤ 1 for any x, then −1/2 < r < 1/2 and so the geometric
2
series converges (for every value of x).
∞
X ∞
X
(e) True/False: If an converges, then a2n converges.
n=1 n=1
∞ ∞
(−1)n X 2
an converges. However, a2
X
Soln: False: If an = √ , then by the alternating series test, the series n = 1/n and so the series an is
n n=1 n=1
the harmonic series which diverges.
∞
bn+1 X 1
(f) True/False: If lim | | = 2, then bn ( )n converges.
n→∞ bn n=1
4
∞
n
X
Soln: True: Consider the power series bn x . Using the ration test to compute the radius of convergence of the power series gives us
n=1

bn+1 xn+1 bn+1

lim | | = |x| lim | | = 2|x|.
n→∞ bn xn n→∞ bn

Thus, by the ratio test, the radius of convergence of this power series is 1/2, that is to say the series converges for −1/2 < x < 1/2. Since x = 1/4
lies in this interval of convergence, the above series converges.
∞
X
(g) If the numbers a0 , a1 , a2 , a3 . . . , decrease to 0, then (−1)n an converges.
n=0
Soln: True: Alternating series theorem.
∞ ∞
X 1 X −1 n
(h) If Cn ( )n converges, then Cn ( ) converges.
n=0
2 n=0
4
∞
n
X
Soln: True: The hypothesis of the statement says that the power series Cn x converges when x = 1/2. But by a theorem about power
n=0
series, the series converges for −1/2 < x < 1/2. In particular, the series converges at x = −1/4.
∞
X
(i) The series (−1.23)n = 1
1+1.23 .
n=0
Soln: False: Although the series is a geometric series with r = −1.23 and the formula is applied correctly, the series does not converge since
|r| > 1. Recall that for a geometric series to converge, the ratio r must satisfy −1 < r < 1.

(k) If the terms of a series alternate, then the series converges.

Soln: False: The series 1 − 1 + 1 − 1 + 1 · · · alternates but does not converge since the terms do not go to zero.
P∞
(l) If an > 0 and an < e−n , then n=1 an converges.
1
2

∞ ∞ ∞
−n −n
X X X
Soln: True: The series e is a convergent geometric series and by hypothesis, an ≤ e .
n=1 n=1 n=1

(m) 1/(1 − x) = 1 + x + x2 + x3 + x4 · · · for every x 6= 1.

Soln: False: This is a geometric series which converges only when −1 < x < 1.

(n) If an > 0 and lim an+1 /an = 1, then the series diverges.
n→∞
Soln: False: Consider the example an = 1/n2 . It is easy to check that the hypothesis are satisfied. However, by the integral test, the series
∞
2
X
1/n converges.
n=1
∞
X
(o) If an > 0 for all even n and an < 0 for all odd n, then an converges.
n=0
∞
Soln: False: If an = (−1)n , then
X
an = 1 − 1 + 1 − 1 + 1 − · · · which diverges.
n=0
∞
X ∞
X
(p) If 0 < an < 1 and an converges, then a2n converges.
n=0 n=0
Soln: True: Since 0 < an < 1, then a2
n < an . Now use the comparison test.

1 1 1
(q) The series 1 − + − + · · · converges.
ln 2 ln 3 ln 4
Soln: True: Alternating series theorem.
∞
X n2
(r) The series converges.
n=1
3n
Soln: True: Ratio test.
1
(s) 1 + e + e2 + e3 + · · · = .
1−e
Soln: False: Can’t use geometric series formula here since e > 1.
∞
X ∞
X
(t) If an xn converges when x = 6, then nan 4n−1 converges.
n=0 n=1
Soln: True: The hypothesis say that the series converges in the interval (−6, 6). The derivative of the power series will also converge in the
same interval. In particular, the derivative of the power series will converge at x = 4.
∞ √
X n
(u) The series converges.
n=1
n+2
√ X √
Soln: False: Compare the terms of the series with 1/ n. The series 1/ n diverges.

(v) 1 − 1 + 1 − 1 + 1 − 1 + · · · = 0.
Soln: False: The series does not converge (partial sums are either 1 or 0) and so it can’t equal anything.

(w) If p is any polynomial, then the Taylor series about any point is equal to p.
Soln: True

2. If an is a sequence of numbers with 0 < an < 1, which of the following conditions will guarantee that the
∞
X
series an converges. Circle as many as necessary. No explanations are required.
n=0

(a) an → 0 as n → ∞.
(b) an ≤ 1/n2 for all n.
∞
X 1
(c) The series converges.
a
n=0 n

(d) an > 1/n for all n

 3

∞
X
(e) The series a2n converges.
n=0
∞
X
Soln: (a) No. Let an = 1/n. Then 0 < an < 1 but the series an = ∞; (b) Yes. Use the comparison and integral tests; (c) No. If the
n=1
∞
X
series 1/an converges, then 1/an → 0 and so an → ∞, which contradicts the hypothesis 0 < an < 1; (d) No. By the comparison test and
n=1
∞ ∞
2
X X
integral test, this says just the opposite; (e) No. If an = 1/n, then an converges. But an does not.
n=1 n=1

∞
X
3. (a) If an → 0, does this mean the series an converges? If it does, show why. If it does not, give an
n=0
example.
∞
X
Soln: No. The series 1/n has terms which do to zero. Yet, by the integral test, does not converge.
n=1

1 1 1 1
(b) What does the following series equal 1 − + − + − · · ·? Hint: Look at ln(1 + x).
2 3 4 5
Soln: The function 1/(1 + x) = 1 − x + x2 − x3 + x4 − · · · for −1 < x < 1. Taking anti-derivatives of both sides of this equation give us

x2 x3 x4
ln(1 + x) = x − + − + ··· .
2 3 4
1 1 1 1
Evaluating at x = 1 gives us ln 2 = 1 − + − + − ··· .
2 3 4 5

(c) Does the series 1 − 1 + 1 − 1 + 1 − 1 + · · · converge? Give a reason to support your answer.
Soln: No. The terms of the series do not go to zero.

(d) What does the series e−1 + e−2 + e−3 + e−4 + · · · converge to?
Soln: This series is a geometric series with r = e−1 (the ratio) and a = e−1 (the first term). Note that −1 < r < 1 and so the series converges

to a/(1 − r) = e−1 /(1 − e−1 ).

∞
1 1 1 1 X 1
(e) Does the series + + + + · · · = converge? Give a reason to support
1 · 2 2 · 22 3 · 23 4 · 24 n=1
n2 n

your answer.
an+1 1
Soln: Yes it does. Let an = 1/n2n and notice that lim | |= < 1. By the ratio test, the series will converge.
n→∞ an 2

4. 1 − 1/3 + 1/9 − 1/27 + · · · =?

Soln: The above us a geometric series with (ration) r = −1/3 and (first term) a = 1. The sum is thus a/(1 − r = 1/(1 + 1/3) = 3/4.

5. Discuss the convergence (or divergence) of the following series.

∞
X 1
(a)
n=2
n log n
∞
X (−1)n
(b)
n=2
n log n
∞
X
(c) (−2)n
n=0

Soln: (a) Diverges (integral test) (b) Converges (alternating series test) (c) Diverges (terms don’t go to zero)

6. Discuss the convergence of divergence of the following series.

∞
X
(a) (.999)n
n=0
∞
X 2
(b) ne−n
n=1
4

∞
X
(c) n10 e−n
n=1

Soln: (a) Converges (geometric series) (b) Converges (integral test - or ratio test) (c) Converges (ratio test)

7. Discuss the convergence (or divergence) of the following series.

(a) 2 − 2 + 2 − 2 + 2 − 2 + 2 · · ·
∞
X 22n
(b)
n=0
(n + 1)!
X∞
(c) 2−3n
n=0
∞
X 1
(d) √
n=0
n log n
−3n n
Soln: (a) Diverges (terms don’t go to zero) (b) Converges (ratio test - with a little algebra) (c) Converges (note that 2 = (1/8) and so
this series is a geometric series) (c) Diverges (integral test)

8. Discuss the convergence (or divergence) of the following series.

∞
X
(a) 2−n
n=0
∞
X 1
(b)
n=2
n ln n
∞
X
(c) n100 e−n
n=0

Soln: (a) Converges (geometric series) (b) Diverges (integral test) (c) Converges (ratio test)

9. Discuss the convergence of the series

∞
X (1 + 1/n)n
.
n=1
2n

(1 + 1/n)n C
Soln: Note that (1 + 1/n)e → e as n → ∞. Thus ' . This last quantity is the term of a geometric series. So, by comparison,
2n 2n
the series converges.

10. Does the following series converge?

∞
X 2n2 + n − 3
n=1
n3 + 2n − 1
∞
2n2 + n − 3 C X
Soln: Note that ' . But the series 1/n diverges (by the integral test) and so, by comparison, the original series diverges.
n3 + 2n − 1 n n=1

11. Does the following series converge? Does it converge absolutely?

∞
X (−1)n n2
n=1
3n

Soln: The series converges since it is an alternating series and the absolute value of the terms monotonically decreases to zero as n → ∞
(alternating series theorem). Taking absolute values of the terms and using the ratio test, one can show that the sum of the absolute values of the
∞
X n2
terms of the series, i.e., converges. Thus the series converges absolutely.
3 n
n=1

12. Does the following series converge? Does it converge absolutely?

∞
X (−1)n
n=1
n log n
 5

∞
X 1
Soln: The series converges, alternating series test, but does not converge absolutely since the series does not converge, by the
n=1 n log n
integral test. So the series converges but not absolutely.

13. Does the following series converge? Does it converge absolutely?

∞
X (−1)n
√
n=1
n
∞
X 1
Soln: The series converges, alternating series test, but does not converge absolutely since the series √ does not converge, by the integral
n=1 n
test. So the series converges but not absolutely.

14. Discuss the convergence (tell if the series converges absolutely, conditionally, or not at all) of the following
series: X
n+1
(a)
X 5n
(b) cos(1/n2 )
X (−1)n n2
(c)
n3 + 2
Soln: (a) Absolutely (ratio test) (b) Diverges (terms don’t go to zero) (c) Conditionally (the series converges - alternating series test - but the

absolute value of the terms is approximately C/n which form the terms of the divergent series - integral test)

15. Compute exactly the following infinite sums. Circle your answer.
(1) 1 − 21 + 13 − 14 + 15 . . .
1 1 1 1
(2) 1 − 2! + 4! − 6! + 8! ...
∞
X2 n
(3)
n=0
n!
∞ ∞
X X 1
(4) e−n = n
n=0 n=0
e

Soln: (1) This series converges to ln 2. The power series for ln(1 + x) = x − x2 /2 + x3 /3−. Now evaluate both sides at x = 1. (2) This series
converges to cos1. See this by writing out the power series expansion of cos x and evaluating both sides at x = 1. (3) This series converges to e2 .
See this by writing out the power series expansion of ex and evaluating both sides at x = 2. (3) This series converges to 1/(1 − 1/e) (geometric

series).

16. A ball is dropped from 1 ft. off the ground. Each time it hits the ground, it bounces 1/3 of the height
it fell. (e.g. On on the first bounce, it rises 1/3 ft.) If the bouncing if allowed to continue forever, find the
total distance travelled by the ball.
1 1 1 1 1 1 1 1 1 1/3
Soln: The distance travelled is 1 + + + + + + + · · · = 1 + 2( + + + ···) = 1 + 2 .
3 3 9 9 27 27 3 9 27 1 − 1/3

17. The bob on a pendulum swings through an arc 24 cm long on its first swing. If each successive swing is
5/6 of the length of the preceding swing, find the total distance the bob travels.
5 5 5 5 5 5 5 5 2 5 3 1
Soln: The total distance travelled is 24 + 24 + 24 + 24 + · · · = 24(1 + +( ) + ( ) + · · · ) = 24 .
6 6 6 6 6 6 6 6 6 1 − 5/6

∞
X an
18. What is e−a ?
n=0
n!
∞ ∞ ∞
a
X an X −a an −a
X an −a a
Soln: Notice that e = and so e =e =e e = 1.
n=0 n! n=0 n! n=0 n!

Z 1
19. Use the power series for sin x to write I = sin xdx as an infinite series. Now use the theory of
0
alternating series to determine how many terms of this series are needed to estimate I within 10−3 . Then
compute this series estimate of I.
6

x3 x5 x7
Z 1
1 1 1 1
Soln: sin x = x − + − + · · · . Thus sin x dx = − + −
+ · · · .. This is an alternating series and so the error is at
3! 5! 7! 0 8! Z 2 4! 6!
1 1 1 1
most the absolute value of the (n + 1)st term. One checks that 1/8! < 10−3 and so sin x dx ≈ − + with an error no more than 10−3 .
0 2 4! 6!

√ √
20. (a) Approximate 1.2 by using the third degree Taylor polynomial (about x = 0) of 1 + x.
2 3
√ x x x
Soln: The third degree Taylor polynomial for this function is 1+x ≈ 1+ − + . Evaluate this polynomial at x = 0.2 to get
√ 2 8 16
1.2 ≈ 1.0955.

(b) Estimate the error.

1 (4) √ 0000 15 1 1 (4)
Soln: The error is at most max |f (x)|, where f (x) = 1 + x. Note that f (x) = − and so max |f (x)| 6
4! 06x60.2 16 (1 + x)7/2 4! 06x60.2
1 15 = 0.0390625.
4! 16

1
21. (a) Compute the power series for
1 + x2
(b) Use this to compute the power series for tan−1 x.
1 1 1 1 1
(c) What is 1 − + − + − + · · ·?
3 5 7 9 11
Soln: The function arctan x is the anti-derivative of 1/(1 + x2 ). The function 1/(1 + x2 ) can be written as
1 2 2 2 2 3 2 4 2 4 6 8
= 1 − (x ) + (x ) − (x ) + (x ) − · · · = 1 − x + x − x + x − · · · .
1 + x2

x3 x5 x7 1 1 1 1 1
Taking anti-derivatives of both sides gives us arctan x = x − + − + · · · . Evaluating at x = 1 gives us 1 − + − + − + ··· =
3 5 7 3 5 7 9 11
arctan 1 = π/4.

22. For what x does the following power series converge (specify where is converges and where it converges
absolutely).
X 4n xn
.
n
4n xn an+1 n
Soln: Let an = . Then = 4|x| . Taking limits to both sides of the above gives us
n an n+1
an+1 n
lim = 4|x| lim = 4|x|.
n→∞ an n→∞ n+1
This limit is less than one when |x| < 1/4. Thus the series converges (absolutely) for −1/4 < x < 1/4. When x = 1/4, the series becomes
∞ ∞
X 1 X (−1)n
which diverges (integral test). When x = −1/4, the series becomes which converges (alternating series test) but not absolutely
n=1 n n=1 n
(integral test).

23. For what x does the following power series converge (specify where is converges and where it converges
absolutely).
X xn
.
2n n
xn an+1
n |x| |x|
Soln: Let an = and note that lim | = |= . Thus, by the ratio test, the series converges absolutely for |x| < 2.
lim
n2n n→∞ 2 n+1an 2 n→∞
∞
X 1
What about the endpoints? When x = 2, the power series becomes which diverges (integral test) while when x = −2, the series becomes
n=1 n
∞
X (−1)n
which converges (alternating series test) but not absolutely (integral test).
n=1 n

∞
X 2n n
24. Compute the radius of convergence of the power series x .
n=1
(n!)2
2n n
Soln: Let an = x and use the ratio test.
(n!)2

an+1 2n+1 xn+1 (n!)2 2

| |=| · |= |x|.
an ((n + 1)!)2 2n xn (n + 1)2
Taking limits as n → ∞ we get this ratio is zero. Thus, by the ratio test, the radius of convergence is infinity. That is to say, the series converges
for all values of x.
 7

25. Find the radius of convergence of the following power series.

∞
X 2n n
(a) x
n=1
(n!)2
2n n |an+1 | 2
Soln: If an = x , then a computation reveals that lim = lim |x| = 0 for all x. Thus, by the ratio test, the series
(n!)2 n→∞ |an | n→∞ (n + 1)2
converges for all x.
∞
X
(b) (n + 1)!xn
n=0
|an+1 |
Soln: If an = (n + 1)!xn , a computation reveals that lim = lim (n + 2)|x| = ∞ for all x. By the ratio test, the series converges
n→∞ |an | n→∞
only for x = 0.

26. The two power series

∞
X ∞
X
an (x − 2)n and bn (x − 3)n
n=0 n=0
both converge when x = 6. Find the largest interval for which both series must converge.
Soln: The first series will converge in the interval (−2, 6]. The second will converge in the interval (0, 6]. The intersection of these two intervals
is (0, 6].

∞
X
27. Suppose f (x) = an xn has a radius of convergence r; i.e., the power series converges to f (x) for
n=0
−r < x < r and does not for |x| > r.
∞
X
What is the radius of convergence for the series f (bx) = an (bx)n ?
n=0
Soln: If the above series has radius of convergence r, then the series f (x) converges for |x| < r and diverges for |x| > r. It follows that the
series f (bx) converges for |bx| < r, i.e., |x| < r/b, and diverges for |bx| > r, i.e., |x| > r/b. Thus the radius of convergence of f (bx) is r/b.
∞
X
What is the radius of convergence for the series f (x2 ) = an x2n ?
n=0
√ √
Soln: By the reasoning above, the series f (x2 ) converges for |x2 | < r, i.e., |x| < r and diverges for |x2 | > r, i.e., |x| > r. Thus the radius
√
of convergence of f (x2 ) is r.

∞
X
28. Suppose that the power series ak (x − 2)k converges when x = 5 and diverges when x = −7. Answer
k=0
the following questions with short answers.
(a) Does the series converge when x = 0?
∞
X
(b) Does the series kak (x − 2)k−1 converge when x = 0?
k=1

(c) Does the series converge when x = 20?

(d) Is the radius of convergence of the series equal to 3?
(e) Does the series converge when x = 11?
Soln: First a general observation. From the theory of power series the interval of convergence must be symmetric about the base point x = 2.
Thus the series converges for −1 < x 6 5 and diverges for x 6 −7 and for x > 11. From here we can get the answers to the questions. (a) Yes;
(b) Yes (this series is the derivative of the given series which will have the same interval of convergence); (c) No; (d) The radius of convergence is
at least 3, but it may be more. Not enough information to tell; (e) Not enough information to tell.

29. Recall that the interval of convergence is the set of all points for which a power series converges.
8

∞
X (−1)n 3n xn
(a) Find the interval of convergence for .
n=1
n4
Soln: Using the ratio test, we see that the series converges when 3|x| < 1, i.e., when −1/3 < x < 1/3. At the endpoint x = 1/3, the series
∞ ∞
X (−1)n X 1
is which is a convergent alternating series. At the other endpoint x = −1/3, the series is which is convergent by the integral
n=1 n4 n=1 n
4

test. Thus the interval of convergence is [−1/3, 1/3].

(b) Find a power series whose interval of convergence is (−1, 1].

∞
X (−1)n xn
Soln: .
n=1 n

∞
X 4n (x − 3)n
30. Determine all x for which the series √ converges.
n=1
n
√
4n (x − 3)n |an+1 | n
Soln: If an = √ , then a computation reveals that lim = lim 4 √ |x − 3| = 4|x − 3|. This is less than one when
n n→∞ |a | n→∞ n +1
n
|x − 3| < 1/4 which holds when 2.75 < x < 3.25. So by the ratio test, the series converges when 2.75 < x < 3.25 and diverges when x > 3.25
∞
X (−1)n
or when x < 2.75. When x = 3.25, the series becomes √ which is convergent (alternating series theorem). When x = 3.25, the series
n=1 n
∞
X 1
becomes √ which diverges (integral test).
n=1 n

1
31. (a) Write down the first 5 terms in the power series expansion of √
1 − x2
1 u 3 2 5 3 35 4
Soln: First notice that √ =1+ + u + u + u + · · ·. Now let u = x2 in the above power series to get
1−u 2 8 16 128
1 1 2 3 4 5 6 35 8
p =1+ x + x + x + x + · · ·.
1 − x2 2 8 16 128

(b) Write down the first 5 terms in the power series expansion of sin−1 x.
−1 1 3 3 5 5 7 35 9
Soln: Integrate both sides of what we just computed to get sin x=x+ x + x + x + x + ···
6 40 112 1152

32. Compute the first 8 terms in the power series expansion of f (x) = tan−1 (x) about x = 0.
1 1 2 3 4
Soln: First note that tan−1 x is the anti-derivative of . Next note that = 1 − u + u − u + u − · · · and so
1 + x2 1+u
1 2 4 6 8
= 1 − x + x − x + x − · · ·.
1 + x2

−1 x3 x5 x7
Now integrate to get tan x=x− + − + · · ·.
3 5 7

33. Compute the first 8 terms in the power series expansion of f (x) = (1 + 2x)−2 about x = 0.
1 d 1 1 2 3
Soln: Notice that =− . Also notice that = 1 − u + u − u + · · · and so
(1 + u)2 du 1 + u 1+u
1 2 3 4
= 1 − 2u + 3u − 4u + 5u − · · · .
(1 + u)2
1 2 3
Finally, setting u = 2x we get = 1 − 4x + 12x − 32x + · · ·.
(1 + 2x)2

34. Compute the Taylor polynomial P3 (x, 0) for the function f (x) = log(1 + x).
1 2 3 4
Soln: First note that = 1 − x + x − x + x + · · · and so, taking anti-derivatives of both sides, we get
1+x
x2 x3 x4
log(1 + x) = x − + − + ··· .
2 3 4

35. Suppose f (x) has power series

∞
1 2 1 3 X 1 n
x + x + x + ··· = 2
x .
4 9 n=1
n
 9

(a) What is f (20) (0)?

f (20) (0) 1
Soln: By the theory of power series, the constant in front of x20 is . In the power series above, the constant in front of x20 is .
20! 202
(20)
f (0) 1 (20) 20!
Thus = for which follows that f (0) = .
20! 202 202

(b) Is f (x) defined at x = 1? Explain.

∞
X 1
Soln: Yes, since, by the integral test, the series converges.
2
n=1 n

(c) Is f (x) defined at x = 2? Explain.

∞
X 2n
Soln: No, since, by the ratio test, the series does not converge.
n=1 n2

36. What degree Taylor polynomial about x = 0 is necessary to approximate cos x in the interval [−2, 2]
with an error less than 10−6 .
Soln: The standard error estimate says that

(N +1) |x|N +1 2N +1
EN (x) ≤ max |f (t)| ≤1· .
0≤t≤x (N + 1)! (N + 1)!

Note that 214 /14! = 1.879 × 10−7 and this is the smallest N that works. Thus we need to take the 13-th degree Taylor polynomial of cos x.

37. (a) Compute the 4th degree Taylor Polynomial of ex .

x x2 x3 x4
Soln: e ≈1+x+ + + .
2! 3! 4!
−.01
(b) Estimate e .
−0.01 (−0.01)2 (−0.01)3 (−0.01)4
Soln: e ≈ 1 + (−0.01) + + + ≈ 0.99005.
2 6 24

(c) Estimate the error.

1 (5) 5 1 5 −13
Soln: If f (x) = ex , the error E4 satisfies the inequality E4 6 max |f (x)||0 − (−0.01)| = (0.01) = 8.333 × 10 .
−0.016x60 5! 5!

(d) How many terms are needed for the error to be less than 10−7 ?
Soln: In general, the error is estimating e−0.01 using an n-th degree Taylor polynomial is (using the a similar calculation as above, is
(0.01)n+1
En 6 . One quickly checks, by trial and error, that when n = 3 that (0.01)n+1 /(n + 1)! < 10−7 .
(n + 1)!

38. Estimate e1/4 using P2 (x, 0) and estimate the error R2 (x, 0).
Soln: This problem is similar to the one above.

39. Estimate sin(1/2) with P5 (x, 0) and estimate the error.

x3 x5 1 31 1 5 11 1
Soln: sin x ≈ x − + and so sin
) +( ) ≈ −(
= 0.479427. Let f (x) = sin x and note that the error E5 satisfies the
6 120 2 6 2 2
120 2
1 (6) 6 1 1 6
inequality E5 6 max |f (x)||0 − 1/2| 6 ( ) = 0.0000217014
6! 06x61/2 6! 2

1
40. Use the fourth degree Taylor polynomial (i.e. P4 ) to approximate √ and estimate the error.
1.2
1 x 3 2 5 3 35 4 1 √
Soln: Note that √ ≈1− + x − x + x . Now evaluate this at x = 0.2 to get ≈ 0.990148. If f (x) = 1/ 1 + x, note
1+x 2 8 16 128 1.2
1 (5) 5 1 945 1 5 1 945 5
that the error E4 is satisfies the inequality E4 6 max |f (0)||0 − 0.2| = max (0.2) = (0.2) = 0.00007875
5! 06x60.2 5! 06x60.2 32 (1 + x)11/2 5! 32

41. Approximate the following integral using power series and explain what you are doing
Z 1/2 p
x3 1 − x3 dx.
0
10

√ u u2 u3 5 4
p x3 x6 x9
Soln: Note that 1−u=1− − − − u − · · · and so 1 − x3 = 1 − − − − · · ·. From this we get that
2 8 16 128 2 8 16

3
p 3 x6 x9 x12
x 1 − x3 = x − − − − ··· .
2 8 16
Thus
(1/2)4 (1/2)7
Z 1/2
3
p
x 1 − x3 dx = − − ··· .
0 4 14

42. How many terms of the power series for ex (about x = 0) are needed to approximate e0.1 to within 10−3 ?
Soln: This problem is similar to one presented earlier.

43. Compute the following limits:

√
cos x − 1
(a) lim
x→0+ x
cos x
(b) lim
x→π/2 (x − π/2)

44. Compute the following limits

2
e−2x − 1
(a) lim
x→0 x2
ex − 1 − x
(b) lim 2
x→0 x cos 2x

45. Using Taylor series, compute the following limits:

4x − sin(4x)
(a) lim
x→0 x3
(sin x − x)3
(b) lim
x→0 x(1 − cos x)4

2
e−2x − 1
46. If f (x) = , compute f (6) (0).
x2
47. (a) Evaluate the following limit
e−2x − 1
lim .
x→0 tan−1 x

−2x 2 4 3 −1 x3 x5
Soln: First note that e = 1 − 2x + 2x − x + · · · and tan x=x− + − · · ·. Thus
3 3 5
e−2x − 1 −2x + 2x2 − 4
3
x3 + · · · x(−2 + 2x − 4 x2 + · · · )
3
−2 + 2x − 4
3
x2 + · · ·
= 3 5
= 2
= .
tan−1 x x x
x − 3 + 5 − ··· x(1 − x3 + x4 − · · · ) 2 4
1 − x3 + x5 − · · ·
5
Taking limits as x → 0 gives us −2.

(b) If f (x) = tan−1 x, what is f (5) (0)?

−1 x3 x5
Soln: Note that tan x=x− + − · · ·. Also note that the coefficient in front of x5 (namely 1/5) is equal to f (5) (0)/5!. Thus
3 5
f (5) (0) = 5!/5 = 4! = 24.

48. If f (x) = cos(x2 ), compute f (100) (0).

49. (a) In planning a highway across a desert, a surveyor must make compensations for the curvature of the
earth when measuring differences in elevation. Using trigonometry one proves that if s is the length of the
highway and R is the radius of the earth, then the correction C is given by
C = R(sec(s/R) − 1).
 11

If f (x) = sec x, the use the following table to get an estimate of C in terms of s and R.
f (0) f 0 (0) f 00 (0) f 000 (0) f 0000 (0)
1 0 1 0 5
(b) Compute the following limit
sec(−2x2 ) − 1
lim .
x→0 x4

50. Suppose we are given a function f (x) such that f (0) = 1, f 0 (0) = 2, f 00 (0) = 3, f 000 (0) = 4, and
f (x) − 1
f (iv) (0) = 5. Compute lim .
x→0 x

51. Find the sum of the following series

∞
X x4n
(a) (−1)n
n=0
n!
∞ n 2n+1
X (−1) π
(b) 2n+1 (2n + 1)!
n=0
4
∞
X (−1)n π 2n
(c)
n=0
62n (2n)!

52. Evaluate exactly the following series.

∞
X 1
(a) n
.
n=0
n!2
∞
n x
X
Soln: This series is equal to x /n! = e when x = 1/2. Thus the series sums to e1/2 .
n=0
∞
X n
(b) n
.
n=1
3
∞
X n−1 1 d 1 1 1
Soln: This series is equal to 1/3 nx = = when x = 1/3. This evaluates to 3/4.
n=1 3 dx 1 − x 3 (1 − x)2