Chapter 10. Sequences, Etc.: Upper Bound

Chapter 10. Sequences, etc.
These typed notes have pictures. Thus you are reminded that it is crucial that you copy good
pictures from class. The most important technique in 10.1–10.4 is to look at the ‘pattern’ you see
emerging in the picture. Actually in 10.2–10.4 the most important technique is to (1) write the
sequence as a long list of numbers, and (2) draw the numbers in (1) as dots on a number line, and
seeing the pattern that is emerging with those dots on the number line.
10.1: Least upper bounds and greatest lower bounds.
• Draw a set S of numbers as a subset of the real number line [picture drawn in class]. An
upper bound of S is a number to the right of S in my picture. [Picture drawn in class.] That
is, an upper bound of S is a number α which is greater than or equal to every number in
S. That is, an upper bound of S is a number α such that x ≤ α for all x in S.
A lower bound of S is a number to the left of S in my picture. [Picture drawn in class.]
That is, a lower bound of S is a number β which is less than or equal to every number in
S. That is, a lower bound of S is a number β such that x ≥ β for all x in S.
• We say that the set S is bounded above, if it has an upper bound. We say that the set S is
bounded below if it has a lower bound. If it is not bounded above, or if it is not bounded
below, then we say that S is unbounded. If it is both bounded above and bounded below,
then we say that S is bounded.
• A maximum of S is an upper bound of S that is in S. A minimum of S is a lower bound
of S that is in S.
• We are mainly interested in this section in the least upper bound of S, written LUB(S) (or
in some books sup(S), and called the supremum). And in the greatest lower bound of S,
written GLB(S) (or in some books inf(S), and called the infimum).
• Viewed in the picture of S on the real number line [picture drawn in class], to find LUB(S)
start at any upper bound to the right of S in the picture, then walk towards S until you
are forced by S to stop. That stopping point is LUB(S). Similarly, to find GLB(S) start
at any lower bound to the left of S in the picture, then walk towards S until you are forced
by S to stop. That stopping point is GLB(S). [Picture drawn in class.]
• Example 1. Find three upper bounds of S = (−∞, 0), and identify the least upper bound
LUB(S).
Solution. [Picture of (−∞, 0) on the real number line drawn in class.] Three upper
bounds are 2, 1, 0. LUB(S) = 0.
• Example 2. Find three lower bounds of S = { 12 , 23 , 34 , 45 , · · · }, and identify the greatest
lower bound GLB(S).
Solution. [Picture of S as dots on the real number line bunching up at 1 drawn in class.]
Three lower bounds are 0, −1, 12 . Also, GLB(S) = 12 .
• Example 3. Determine if the set S = [0, 4] is bounded above. If it is, determine the least
upper bound LUB(S).
Solution. [Picture of S as an interval on the real number line drawn in class.] Yes, S is
bounded above, and LUB(S) = 4.
• Example 4. Determine if the set S in Example 2 is bounded above. If it is, determine the
least upper bound LUB(S).
Solution. Look at the picture of S drawn in Example 2. Yes, S is bounded above, and
LUB(S) = 4.
• Example 5. Determine if the set S = {x : x2 < 9} is bounded above. If it is, determine
the least upper bound LUB(S).
Solution. [Picture of S as the interval (−3, 3) on the real number line drawn in class.]
Yes, S is bounded above, and LUB(S) = 3.
• Example 6. Determine if the set S = {x : |x− 2| < 3} is bounded above. If it is, determine
Solution. [Picture of S as the interval (−1, 5) on the real number line drawn in class.]
Yes, S is bounded above, and LUB(S) = 5.
• Example 7. Determine if the set S = {x : ln x < 2} is bounded above. If it is, determine
Solution. ln x < 2 is the same as x < e2 . Note that x > 0 since ln x is not defined for
negative numbers. [Picture of S as the interval (0, e2 ) on the real number line drawn in
class.] Yes, S is bounded above, and LUB(S) = e2 .
• Example. Redo Examples 1–7, but with ‘bounded above’ replaced by ‘bounded below’
and least upper bound LUB replaced by greatest lower bound GLB.
Solutions. 1) is not bounded above, so no greatest lower bound or GLB. We did 2) and
4) already. The set in 3) is bounded below and GLB(S) = 0. The set in 5) is bounded
below and GLB(S) = −3. The set in 6) is bounded below and GLB(S) = −1. The set in
7) is bounded below and GLB(S) = 0.
10.2: Sequences.
• A sequence is a numbered string of objects a1 , a2 , a3 , · · · . We often write the sequence as

(an )∞
n=1 or simply as (an ). The nth term is an . In this course our sequences are infinite
sequences of real numbers.
• Example 1. List (n2 − 2n + 1)∞ n=1 as a string of numbers.

Solution. 0, 1, 4, 9, · · · .
• Example 2. List (n2 − 2n + 1)∞ 2 4

n=2 and (n − 2n + 1)n=2 , as strings of numbers (the second
is a finite sequence.
Solution. 1, 4, 9, 16, · · · ; and 1, 4, 9.
n
• Example 3. List (1 + (−1) n ) as a string of numbers, and draw them on the real number line
noting the pattern that emerges.
Solution. 0, 1 + −1 1 −1 1 3 2 5
1 , 1 + 2 , 1 + 3 , 1 + 4 , · · · . This is 0, 2 , 3 , 4 , · · · . [Picture drawn in class
on the real number line of dots alternating between being to the left and to the right of
1, but always getting closer and closer to 1. These dots each get a number above them: 1
above 0, write 2 above 32 , write 3 above 32 , and so on.]
The technique in Examples 1–3 you will use in almost every problem: listing a sequence as a
string of numbers, and drawing them on the real number line noting the pattern that emerges.
A question you will often encounter in the quizzes and tests is to do the opposite of the procedure
we used in Examples 1–3. In each of the examples below the trick I taught in class is to write 1
above the first term in the sequence, 2 above the second term in the sequence, 3 above the third
term in the sequence, and so on, and then find how each number 1, 2, 3, · · · is connected to the
number below it.
• Example 4. Assuming that the pattern continues as indicated, find an explicit formula for
the nth term an : Here the sequence is −1, 1, −1, 1, 1, 1, −1, · · · .
Solution. an = (−1)n . So the sequence is ((−1)n ).
the nth term an : Here the sequence is 1, 8, 27, · · · .
Solution. an = n3 . So the sequence is (n3 ).
the nth term an : Here the sequence is 21 , 23 , 34 , 45 , · · · .
n n
Solution. an = n+1 . So the sequence is ( n+1 ).
the nth term an : Here the sequence is 0, 34 , 45 , 87 , 89 , 12 12
11 , 13 , · · · .
2n−1+(−1)n (−1)n (−1)n
Solution. an = 2n−1 =1+ 2n−1 . So the sequence is (1 + 2n−1 ).
the nth term an : Here the sequence is 23 , − 94 , 27 81
8 , − 16 , · · · .
n
Solution. an = (−1)n+1 32n = −(− 32 )n . So the sequence is (−(− 32 )n ).
A sequence is called monotone if it is either increasing or decreasing. The boundedness and

monotonicity of a sequence can almost always be discovered by the procedure we did above of
listing a sequence as a string of numbers, and drawing them on the real number line noting the
pattern that emerges.
Example. Determine the boundedness and monotonicity of each of the following sequences: (1)
(−1)n
the sequence (n2 − 2n + 1)∞ 2 ∞
n=1 , (2) the sequence (n − 2n + 1)n=2 , (3) the sequence (1 + n ),
n 3 1 2 3 4
(4) the sequence ((−1) ), (5) the sequence (n ), (6) the sequence 2 , 3 , 4 , 5 , · · · , (7) the sequence
n
(1 + (−1) 3 n
2n−1 ), (8) the sequence (−(− 2 ) ).
Solution. These are just the 8 sequences we looked at a few moments ago. The sequence in
Example (1) is 0, 1, 4, 9, · · · . When one looks at this as dots on the number line at the points
0, 1, 4, 9, and so on, [Picture drawn in class], one sees that Example (1) is bounded below, but not
above (if you are asked for it, the GLB is 0). It is increasing.
The sequence in Example (2) is 1, 4, 9, · · · . When one looks at this as dots on the number line
at the points 1, 4, 9, and so on, one sees that Example (2) is bounded below, but not above (if you
are asked for it, the GLB is 1). It is increasing.
The sequence in Example (3) we saw above is the same as 0, 23 , 23 , 54 , · · · . [Picture drawn in class
on the real number line of dots alternating between being to the left and to the right of 1, but
always getting closer and closer to 1. These dots each get a number above them: 1 above 0, write 2
above 32 , write 3 above 23 , and so on.] This sequence is bounded below and bounded above, so it is
bounded (if you are asked for it, the GLB is 0 and the LUB is 23 ). This sequence is not monotonic.
The sequence in Example (4) is the same as −1, 1, −1, 1, 1, 1, −1, · · · . This sequence is bounded
below and bounded above, so it is bounded (if you are asked for it, the GLB is −1 and the LUB is
1). This sequence is not monotonic.
The sequence in Example (5) is 1, 8, 27, · · · . When one looks at this as dots on the number line
at the points 1, 8, 27, and so on, one sees that Example (5) is bounded below, but not above (if you
are asked for it, the GLB is 1). It is increasing.
The sequence in Example (6) is 12 , 23 , 34 , 54 , · · · . [Picture drawn in class on the real number line
of dots between 21 and 1 but getting closer and closer to 1. These dots each get a number above
them: 1 above 12 , write 2 above 32 , write 3 above 45 , and so on.] This sequence is bounded below
and bounded above, so it is bounded. It is increasing.
The sequence in Example (7) is 0, 43 , 45 , 87 , 89 , 12 12
11 , 13 , · · · . [Picture drawn in class on the real number
line of dots between 0 and 2, alternating between being to the left and to the right of 1, but always
getting closer and closer to 1. These dots each get a number above them: 1 above 0, write 2 above
4 4
3 , write 3 above 5 , and so on.] This sequence is bounded below and bounded above, so it is
bounded. This sequence is not monotonic.
The sequence in Example (8) is 32 , − 49 , 27 81
8 , − 16 , · · · . [Picture drawn in class on the real number
line of dots alternating between being to the left and to the right of 0, but always larger and larger
in absolute value. These dots each get a number above them: 1 above 32 , write 2 above − 94 , write 3
above 278 , and so on.] This sequence is neither bounded below nor bounded above. This sequence
is not monotonic.
The following are done like the last examples, but are a bit harder (except for the first):
q
Example 9. Determine the boundedness and monotonicity of the sequence ( 1 − n1 ).
q q q q
Solution. This is the sequence 12 , 23 , 34 , 45 , · · · , which is ‘the square root’ of the sequence
in Example (6) above. The picture will be similar to the picture drawn in class on the real number
line for Example (6), but with the dots slightly to the right of where they were before. Namely,
dots between 12 and 1 but getting closer and closer to 1. These dots each get a number above them:
q q
1 above 12 , write 2 above 23 , and so on.] This sequence is bounded below and bounded above,
so it is bounded. It is increasing.
Example 10. a) Write down the first 6 terms, and then the general term of the sequence
defined recursively (or inductively) by a1 = 1, an+1 = a2n + 1.
b) Determine the boundedness and monotonicity of the sequence in (a).
Solution. (a) This is the sequence a1 = 1, a2 = a21 + 1 = 12 + 1 = 23 , and
3
a2 7
a3 = +1= 2 +1 = ,
2 2 4
7
a3 15
a4 = +1= 4 +1= ,
2 2 8
15
a4 31
a5 = +1= 8 +1= ,
2 2 16
31
a5 63
a6 = + 1 = 16 + 1 = .
2 2 32
n
The general term an = 22n−1−1
, as one can see similarly to how we proceeded when we first met
Examples 4–8 at the start of Section 10.2 above.
(b) [Picture drawn in class on the real number line of dots between 1 and 2, These dots each
get a number above them: 1 above 1, write 2 above 32 , write 3 above 74 , and so on.] This sequence
is bounded below (by 1) and is bounded above (by 2). This sequence is increasing.
Example 11. Determine the boundedness and monotonicity of the sequence ( enn ).
Solution. Method 1: On a calculator compute e11 , e22 , e33 , · · · , draw these on the real number line,
and note the pattern that emerges. One sees that the sequence is decreasing and bounded below
(by 0) and bounded above (by 1e ).
Method 2: Let f (x) = exx . We sketched the graph of this at the end of Section 7.4 (see typed
notes), and saw that f is decreasing for x ≥ 1. So ( enn ) is decreasing. These numbers are positive,
so the sequence is bounded below by 0 and bounded above its first term 1e since it is decreasing
[Picture drawn in class on the real number line of dots on the graph of y = exx decreasing in height.]
1
Example 12. Determine the boundedness and monotonicity of the sequence (n n ).
1 1 √ √ √
Solution. This is the sequence 1, 2 2 , 3 3 , · · · , which is 1, 2, 3 3, 4 4, · · · .
Method 1: On a calculator compute these numbers, draw these on the real number line, and
note the pattern that emerges. The first two or three terms are increasing, but then the sequence
starts to steadily decrease, but is always bigger than 1. So it is not monotone, is bounded below
(by 1), and is bounded above. [Picture drawn in class on the real number line of dots]
1 1
Method 2: Let f (x) = x x = e x ln x . If one computes f ′ (x) by logarithmic differentiation one
1 √ √
gets f ′ (x) = x x 1−ln
2
x√
x
. This is negative if x > e. So f is decreasing for x ≥ e. So 3
3, 4
4, · · · is
decreasing. But 1, 2 is increasing, so the sequence is not monotone. These numbers are positive,
so the sequence is bounded below by 0. The sequence is bounded above, because after the 3rd term
it is decreasing and bigger than 0 [Draw a picture].
n 2
Example 13. Determine the boundedness and monotonicity of the sequence ( n+3 ).
n2 1 1 3
Solution. Divide top and bottom by n2 : n+3 = 1
+ 3
. Both n and n2 are decreasing, so the
n n2
denominator n1 + n32 is decreasing and positive. So the sequence is increasing. These numbers are
positive, so the sequence is bounded below by 0. The denominator n1 + n32 gets arbitrarily close to
0, so the sequence gets arbitrarily large. So the sequence is unbounded above.
Another way to see that the sequence is unbounded above is to look at the ‘winning term’ in the
n2
numerator and ‘winning term’ in the denominator and divide them. So, for example n+3 acts like
n2
n = n, and the latter is unbounded above. Another way to see that the sequence is monotone:
x2 2x(x+3)−x2 x2 +6x
let f (x) = x+3 , then f ′ (x) = (x+3)2 = (x+3)2 > 0 for x > 0. So f is increasing, and hence so is
n2
f (n) = n+3 .
10.3: Limits of sequences.

• We say that a sequence (an ) converges to a real number a if the numbers an are getting
closer to closer to a, as close as we like, as n gets huge. In this case, we write a = limn→∞ an ,
or an → a as n → ∞. We also say an approaches a as n → ∞.
• The technical definition: a = limn→∞ an if for all numbers ǫ > 0, there exists a number N
such that |an − a| < ǫ whenever n ≥ N .
• Note that a = limn→∞ an if and only if limn→∞ (an − a) = 0.
• If (an ) does not converge to any number then we say that it diverges.
• If the terms in the sequence are all positive and are getting huge without any bound,
we write limn→∞ an = ∞. Note that if the terms in the sequence are all positive then
limn→∞ an = ∞ if and only if limn→∞ a1n = 0. Similarly, if the terms in a sequence are all
negative then limn→∞ an = −∞ if and only if limn→∞ a1n = 0.
• Suppose that (an ) is increasing. Then limn→∞ an is the same as the LUB (and equals ∞ if
(an ) is not bounded).
Suppose that (an ) is decreasing. Then limn→∞ an is the same as the GLB (and equals
−∞ if (an ) is not bounded).
• FACT: A convergent sequence is bounded. Hence:
An unbounded sequence must be divergent.
• (Warning: a bounded sequence does not HAVE to be convergent. See, for example, Example
4 in Section 10.2.)
• Example. State if the following sequences converge or diverge. If the sequence converges,
find its limit. If it diverges, explain why. (1) the sequence (n2 − 2n + 1)∞ n=1 , (2) the
2 ∞ (−1)n
sequence (n − 2n + 1)n=2 , (3) the sequence (1 + n ), (4) the sequence ((−1)n ), (5)
n
the sequence (n3 ), (6) the sequence 12 , 23 , 43 , 45 , · · · , (7) the sequence (1 + (−1)
2n−1 ), (8) the
q
sequence (−(− 32 )n ), (9) the sequence ( 1 − n1 ).
Solution. These are just the 9 sequences we looked at at the start of Section 10.2. The
sequence in Example (1) is 0, 1, 4, 9, · · · . Since this is not bounded, but is increasing, the
limit of the sequence is ∞, by the third last bullet above. The sequence diverges because it
is unbounded (see second last bullet above). Similarly, the limit of the sequence in Example
(2) is ∞, and this sequence diverges.
The sequence in Example (3) converges, and its limit is 1 (since we said when we looked
at Example (3) at the start of Section 10.2, or its picture that we drew there, that these
numbers are getting closer and closer to 1.
The sequence in Example (4) diverges, since these numbers are not getting closer and
closer to any one number.
The sequence in Example (5) diverges (the reasoning is identical to that of Example 1 a
few lines above).
The sequences in Example (6), (7), and (9) converge, and their limit is 1 (since we said
when we looked at those examples at the start of Section 10.2, or their pictures that we
drew there, that these numbers are getting closer and closer to 1.
The sequence in Example (8) diverges, since as we said when we looked at Example (8)
at the start of Section 10.2, it is unbounded. But we said recently that every unbounded
sequence diverges.
an
• Example. Consider the sequence defined recursively by a1 = 1, an+1 = 2 + 1. State if the
sequences converges or diverges. If it converges, find limn→∞ an .
Solution. This is Example (10) at the start of Section 10.2, and we said there, or by the
pictures that we drew there, this is an increasing sequence which is getting closer and closer
to 2. So limn→∞ an = 2.
• Example. State if the sequence (sin( n1 )) converges or diverges. If it converges, find its limit.
1
Solution. It converges with limit 0. To see this, note that as n gets huge, n → 0. Because
sin is continuous, sin( n1 ) is approaching sin 0 = 0.
nπ
• Example. State if the sequence (cos( n+1 )) converges or diverges. If it converges, find its
limit.
nπ
Solution. It converges with limit −1. To see this, note that as n gets huge, n+1 → π.
nπ
Because cos is continuous, cos( n+1 ) is approaching cos π = −1.
• Example. Does the following sequence converge:
1 3 1 5 3 1 7 5 3 1
, · , · · , · · · , ··· .
2 4 2 6 4 2 8 6 4 2
Solution. This is a nonnegative sequence, and notice that each term in this sequence is
equal to its predecessor multiplied by a positive number less than 1. Thus this sequence is
decreasing. By the bullet before the FACT on the last page, this sequence converges to its
greatest lower bound.
• A very common trick to find limits of sequences is to divide numerator and denominator
through by the highest power in the denominator. We do several examples of this:
• Example. State if the sequence ( 5n+1
2n−3 ) converges or diverges. If it converges, find its limit.
1
5n+1 5+ n
Solution. Divide numerator and denominator through by n. We get 2n−3 = 3 ,
2− n
which
5+0 5 5
has limit 2−0 = 2. So the sequence converges with limit 2.
2
• Example. State if the sequence ( 3n +1
n3 −1 ) converges or diverges. If it converges, find its limit.
3
3n2 +1 + 1
Solution. Divide numerator and denominator through by n3 . We get n3 −1 = n n3
1− 13
,
n
0+0
which has limit 1−0 = 0. So the sequence converges with limit 0.
n
• Example. State if the sequence ( √n+1 ) converges or diverges. If it converges, find its limit.
√
Solution. Divide numerator and denominator through by n. We get
√n √
n n n
√ = √ =q .
n+1 n+1
√ 1 + 1
n n
√
The numerator here goes to ∞ and the denominator goes to 1 + 0 = 1. So the sequence
diverges, and its limit is ∞.
• ‘Cheat’ in the previous examples: look at ‘winning term’ in the numerator and ‘winning
term’ in the denominator and divide them. So, for example 5n+1 5n 5
2n−3 acts like 2n = 2 ; so the
2 +1 3n2 √
limit is 52 . And 3n 3
n3 −1 acts like n3 = n which has limit 0. And
√n
n+1
acts like √nn = n
which has limit ∞; so the limit is ∞.
• One can turn many limits (including most of the ones above) into a Calculus I problem by
the following principle: If limx→∞ f (x) = L in the Calculus I sense, then limn→∞ an = L,
if an = f (n).
• Example. State if the sequence (( 23 )n ) converges or diverges. If it converges, find its limit.
Solution. Let f (x) = ( 32 )x . We saw the graph of f in Chapter 7 [Picture drawn in class.]
So limx→∞ f (x) = 0, hence limn→∞ ( 23 )n = 0. So the sequence converges with limit 0.
• Example. State if the sequence ( lnnn ) converges or diverges. If it converges, find its limit.
ln x ∞
Solution. The answer will be the same as limx→∞ x . This limit is of form ∞. By
L’Hospitals rule (later),
d 1
ln x dx (ln x) x
lim = lim d
= lim = 0.
x→∞ x x→∞
dx x
x→∞ 1
So the sequence converges with limit 0.
n
• Example. State if the sequence ( 3n2+1 ) converges or diverges. If it converges, find its limit.
2n ( 23 )n
Solution. Divide numerator and denominator through by 3n . We get 3n +1 = 3+ 31n
, which
0
has limit 3+0= 0 (see previous problem). So the sequence converges with limit 0. Quicker:
look at the ‘winning term’ in the numerator and ‘winning term’ in the denominator and
divide them. One gets ( 32 )n , which has limit 0.
• Example. State if the sequence (ln(n + 1) − ln n) converges or diverges. If it converges, find
its limit.
1
Solution. ln(n + 1) − ln n = ln n+1
n = ln(1 + n ). This has limit ln 1 = 0. So the sequence
converges with limit 0.
• It is sometimes helpful to remember that whether a sequence converges or diverges, has noth-
ing to do with its first few terms. Thus, for example, the sequence 500, 1000, 1500, 2000, 1, 12 , 13 , 14 , 15 , · · · ,
converges to 0.
• The ‘squeezing’ or ‘pinching rule’: uppose that (sn ), (xn ), and (tn ) are sequences with
sn ≤ xn ≤ tn , for every n ≥ 1. If limn sn = s and limn tn = s, then limn xn = s.
• If limn sn = s and limn tn = t, then:
(1) limn sn + tn = s + t;
(2) limn sn − tn = s − t;
(3) limn sn tn = st;
(4) limn Csn = Cs, if C is a constant;
(5) limn stnn = st , if t 6= 0;
(6) limn |s
√n | = |s|√;
(7) limn tn = t, if tn ≥ 0 for all n ∈ N.
10.4: Some important limits.

1 1 1 1
• x n → 1 as n → ∞ if x > 0 (since x n = e n ln x , and ln x
n → 0, so e n ln x → e0 = 1.)
• xn → 0 as n → ∞ if |x| < 1 (this is proved like in the example in Section 10.3 of ( 32 )n ).

• (xn ) diverges if x > 1 or x ≤ −1 (since these cases are unbounded, or is Example (4) in
Section 10.3).
1 1
• np → 0 as n → ∞ if p > 0 (since np = ( n1 )p and 1
n → 0).
• Variant of the ‘squeezing’ or ‘pinching’ rule: Suppose that 0 ≤ an ≤ bn . If bn → 0 then
an → 0. If an → ∞ then bn → ∞.
• Example. State if the sequence ( sinn n ) converges or diverges. If the sequence converges, find
its limit.
Solution: Note that −1 ≤ sin n ≤ 1, so − n1 ≤ sinn n ≤ n1 . But n1 → 0, so by ‘squeezing’
( sinn n ) converges with limit 0.
• We give another proof, using ‘squeezing’, that ( lnnn ) converges to 0. We first notice that
√ 1 1
√
x−2
ln x ≤ x for x ≥ 4, since by Calculus if f (x) = x 2 −ln x then f ′ (x) = 12 x− 2 − x1 = 2x >0
if x > 4. So f (x) is strictly increasing on [4, ∞), and since f (4) = 2 − ln √ 4 > 0, we must
√ √
have f (x) = x − ln x ≥ 0 for x ≥ 4. So x ≥ ln x for x ≥ 4. Thus lnnn ≤ nn = √1n . Since
√1 → 0 by ‘squeezing’ we must have ln n
→ 0.
n n
√
• In the last item we showed that ln x ≤ x. In fact ln x ≤ xp for any p > 0 (the proof is
similar). Similarly, xp ≤ ax for any p and any a > 1, at least for large x.
1
• Example. State if the sequence (n n ) converges or diverges. If the sequence converges, find
its limit.
1 1
Solution: n n == e n ln n → e0 = 1 as n → ∞, since we just saw that lnnn → 0.
2n
• Example. Show that n! → 0 as n → ∞.
2n
Solution: 0 ≤ = n2 n−1
n!
2 2 22
n−2 · · · 2 1 ≤
2
n · 1 · 1··· · 1 · 2 = 4
n. But 4
n → 0, so by by
n
‘squeezing’ we must have 2n! → 0.
xn
• Similarly, n! → 0 as n → ∞.
• (1 + nx )n → ex as n → ∞. We shall do this as an application of L’Hospitals rule in the next
section.
10.5–10.6: L’Hospitals rule.

x
• We have already used L’Hospitals rule twice in this class. First we looked at limx→∞ ex ,
which is called an indeterminant form of type ∞
∞ . We saw that
d
x dx x 1
lim = lim d x
= lim = 0.
x→∞ ex x→∞ dx e
x→∞ ex
ln x ∞
Similarly, limx→∞ x is an indeterminant form of type ∞, and
d 1
ln x dx (ln x) x
lim = lim d
= lim = 0.
x→∞ x x→∞
dx x
x→∞ 1
sin x
• Example. Find limx→0 x .
Solution. This is an indeterminant form of type 00 . By L’Hospitals rule
sin x cos x
lim = lim = cos 0 = 1.
x→0 x x→0 1
ex
• Example. Find limx→∞ x.
ex
Solution. By L’Hospitals rule = limx→∞ 1 = ∞.
ln x
• Example. limx→1 x−1 is an indeterminant form of type 00 . By L’Hospitals rule
d 1
ln x dx (ln x) x
lim = lim d
= lim = 1.
x→1 x − 1 x→1 dx (x − 1)
x→1 1
ln x
• Example. Find limx→0+ 1 .
ex
Solution. This is an indeterminant form of type −∞ ∞ (actually we still call this an inde-
terminant form of type ∞ ∞ ), and by L’Hospitals rule it equals
d 1
dx (ln x) x −x 1
lim 1 = lim 1 = lim 1 = −( lim x) ( lim 1 = 0.
x→0+ d
dx (e )
x x→0+ e x (− x12 ) x→0+ e x x→0+ x→0+ ex
• These kinds of examples can always be applied to sequences (as we did in Section 10.3 when
we showed limn→∞ lnnn = 0). Here is another example of this:
ln n
• Example. Find limn→∞ en .
Solution. Set x = n1 . As n → ∞ we have x → 0+ , so by the previous example
− ln n ln n1 ln x
n
= n
= 1 → 0.
e e ex
ln n
So limn→∞ en = 0.
sin x−x
Example. Find limx→0 x3
.
Solution. This is an indeterminant form of type 00 . By L’Hospitals rule it equals
limx→0 cos3xx−1
2 , which is again an indeterminant form of type 00 . By L’Hospitals rule it
equals limx→0 − 6xsin x
= − 16 .
• Example. Find limx→0+ x ln x.
ln x
Solution. limx→0+ x ln x = limx→0+ 1 , and the latter is an indeterminant form of type
x
−∞ ∞
∞ (actually we still call this an indeterminant form of type ∞ ). By L’Hospitals rule it
1
equals lim x
x→0+ − 1 = − limx→0+ x = 0.
x2
• Example. Find limx→0+ xx .

Solution. This is called an indeterminant form of type 00 . Note that xx = ex ln x → e0 = 1
as x → 0+ , since we just saw that limx→0+ x ln x = 0.
• Example. Find limx→∞ x1 0x 1+t1 R
dt.
Solution. By L’Hospitals rule it equals
d Rx 1 1
dx 0 1+t 1+x
lim d
= lim = 0.
x→∞
dx x
x→∞ 1
• Example. Show that limx→∞ (1 + xa )x = ea , and that limx→∞ (1 + x1 )x = e.

a a
a ln(1+ x ) ln(1+ x )
Solution. (1 + xa )x = ex ln(1+ x ) . Now x ln(1 + xa ) = 1 , and limx→∞ 1 is an
x x
indeterminant form of type 00 . By L’Hospitals rule
1
ln(1 + xa ) a
1+ x (− xa2 ) a
lim 1 = lim = lim a = a.
x→∞
x
x→∞ − x12 x→∞ 1+ x
a
Thus as x → ∞ we have that x ln(1 + xa ) → a, and so (1 + xa )x = ex ln(1+ x ) → ea .
For the last assertion set a = 1.
(x)
• Proof of L’Hospitals rule: We just do the case that limx→c fg(x) is an indeterminant form
0
of type 0 . So limx→c f (x) = limx→c g(x) = 0. We assume that f and g, and f ′ and g′ , are
continuous at c. Then f (c) = g(c) = 0 and
f (x)−f (c)
f (x) f (x) − f (c) x−c f ′ (c) f ′ (x)
lim = lim = lim = = lim .
x→c g(x) x→c g(x) − g(c) x→c g(x)−g(c) g′ (c) x→c g′ (x)
x−c
0 ∞
• L’Hospitals rule ‘does not work’ if the limit is not of form 0 or ∞.
10.7: Improper integrals.

• There are three basic kinds of improper integrals. To describe them recall that in an
integral ab f (x) dx, we call a and b the lower and upper limits of integration, and f (x) is
R
the integrand. The first kind of improper integral is when a = −∞ or b = ∞ or both. The
second kind of improper integral is when a or b are numbers that f (x) ‘blows up’ at. For
example, 01 √1x dx, note that √1x is not defined at 0, and goes to infinity as x → 0+ . The
R
third kind of improper integral ab f (x) dx is when f (x) ‘blows up’ at a number between a
R
R1 1 1
and b. For example, −2 4 dx. Here 4 ‘blows up’ at 0, and 0 is between −2 and 1.
x5 x5 R∞ dx
An integral may be more than one of the above types, like −∞ x2 dx.
• Example. Determine whether or not the integral 1∞ dx

R
x2 dx is improper. Give a reason for
your answer. Also, by taking an appropriate limit, compute the value of the integral.
Solution. This is an improper integral of the first kind, since one of the limits of integra-
tion is ∞. Also,
Z ∞ Z C
dx dx
2
dx = lim dx.
1 x C→∞ 1 x2
Now Z C
dx 1
2
= −x−1 ]C
1 =− +1→1
1 x C
as C → ∞. We say that this improper integral converges and 1∞ dx
R
x2
dx = 1.
1
The last example is really saying that the area underneath the graph of y = x2 for x ≥ 1,
is 1. [Picture drawn in class.]
• Example. Determine whether or not the integral 1∞ dx

R
x dx is improper. Give a reason for
your answer. Also, by taking an appropriate limit, compute the value of the integral.
Solution. ThisRis an improper integral of the first kind, since one of the limits of integra-
tion is ∞. Also, 1∞ dx
R C dx
x dx = lim C→∞ 1 x dx. Now
Z C dx
= ln x]C
1 = ln C − ln 1 = ln C → ∞
1 x
R∞ dx
as C → ∞. We say that this improper integral diverges and 1 x dx = ∞.
1
The last example is really saying that the area underneath the graph of y = x for x ≥ 1,
is 1. [Picture drawn in class.]
R∞ dx R∞ dx 1
• Similarly 1 xp diverges (= ∞) if 0 ≤ p ≤ 1, and 1 xp = p−1 (it converges) if p > 1.
R0 dx R∞ dx
• Compute the following improper integrals if they converge: (a) −∞ x2 +1 , and (b) −∞ x2 +1 .
R0 dx
Solution. (a) The first integral is limC→−∞ C x2 +1 . Now
Z 0 dx −π π
= tan−1 x]0C = tan−1 0 − tan−1 C = − tan−1 C → −( )=
C x2 + 1 2 2
0
as C → −∞. [Drew picture of the tan−1 graph in class.] So −∞ dx π R
R∞ x2 +1 = 2 .
0 ∞ dx ∞ dx
(b) Also, −∞ x2dx dx
R R R
+1 = −∞ x2 +1 + 0 x2 +1 . To do 0 x2 +1 we observe that
Z C dx π
= tan−1 x]C
0 = tan
−1
C − tan−1 0 = tan−1 C →
0 x2 +1 2
R∞
as C → ∞ [Drew picture of the tan−1 graph in class]. So 0
dx
x2 +1 = π2 , so that
Z ∞ dx
Z 0 dx
Z ∞ dx π π
2
= 2
+ = + = π.
−∞ x +1 −∞ x +1 0 x2 +1 2 2
• Example. Determine whether or not the integral 01 √ dx R

x
is improper. Give a reason for your
answer. Also, by taking an appropriate limit, compute the value of the integral.
Solution. This is an improper integral of the second kind, since √1x blows up at 0. The
R1 dx
integral is limC→0+ C
√
x
. Now
Z 1 dx 1 1
√ = 2x 2 ]1C = 2 − 2C 2 → 2
C x
R1
as C → 0+ . So the improper integral converges, and 0
dx
√
x
= 2.
• Example. Determine whether or not the integral 01 ( √1x − (x−1)1 R

2 ) dx is improper. Give a
reason for your answer. Also, compute the value of the integral.
1
Solution. This is an improper integral of the second kind, since ( √1x − (x−1)2 ) blows up
both at 0 and at 1. Let us write

1
Z 1 1 1
Z
2 1 1
Z 1 1 1
(√ − ) dx = (√ − ) dx + (√ − ) dx.
0 x (x − 1)2 0 x (x − 1)2 1
2
x (x − 1)2
1 1
1 2
The first integral is limC→0+ (2x 2 + x−1 )]C , which is a finite number. However the second
integral is
1 1 1 1 √
lim (2x 2 + )]C1 = lim (2C 2 + ) − ( 2 − 2) = −∞,
C→1− x−1 2 C→1 − C −1
1 R1
since C−1 → −∞ as C → 1− . So the integral diverges and 0 ( √1x − 1
(x−1)2
) dx is a finite
number plus −∞, which is −∞.
R1 1
• Example. Determine whether or not the integral −2 4 dx is improper. Give a reason for
x5
your answer. Also, compute the value of the integral.
1
Solution. This is an improper integral because 4 ‘blows up’ at 0, and 0 is between −2
R1 R0 R1 x5
1 1 1
and 1. We write −2 4 dx = −2 4 dx + 0 4 dx, a sum of two of the second kind of
x5 x5 x5
improper integral. To do the first integral here,
Z C
1 1
C 1 1 1 1 1
4 dx = 5x ]−2 = 5C
5 5 − 5(−2) 5 = 5C 5 + 5 · 2 5 → 5 · 2 5 ,
−2 x 5
R0 1
1
as C → 0− . So −2 4 dx = 5 · 2 5 . To do the second integral,
x5
Z 1 1 1 1
4 dx = 5x 5 ]1C = 5 − 5C 5 → 5
x C 5
R1 1
as C → 0+ . So 0 4 dx = 5. Thus
x5
Z 1 1
Z 0 1
Z 1 1 1
4 dx = 4 dx + 4 dx = 5 · 2 5 + 5.
−2 x 5 −2 x 5 0 x 5
• A ”Comparison test”: If 0 ≤ f (x) ≤ g(x) on my interval, and if the integral of g on the

interval converges then the integral of f converges too. If the integral of f on the interval
diverges then the integral of g diverges too.
R∞ R∞
• Do the integrals converge or diverge: (a) √ dx dx, (b) √ dx .
1 1+x2 1 1+x3
Solution. (b) √ 1 < √1 = 1
3 if x ≥ 1. Now since p = 3
> 1, we know from the last
1+x3 x3 2
R∞ x2 R∞
dx √ dx
page that 1 3 converges. So by the ‘comparison test’ above, converges. 1 1+x3
x2
√
(a) √ 1 > 2x1
if x ≥ 1 (to see this note that this is saying that 2x > 1 + x2 or
1+x2
4x2 > > 1, which is true if x ≥ 1). However 1∞ dx
1 + x2 or 3x2
R
R ∞ dx 2x diverges as we saw on the
last page. So by the ‘comparison test’ above, 1 √1+x2 diverges too.
• Gabriel’s horn (textbook p. 624). In the bible, Gabriel is an archangel, and the horn is
the trumpet blown at the end of time, heralding a cataclysmic event described in the book
of Revelation. In this problem, we obtain this infinite trumpet by taking the graph of
f (x) = x1 for 1 ≤ x < ∞, and rotating it around the x-axis [Picture drawn in class.] The
volume formula for this from Calc I is
Z ∞ Z ∞
2 dx
V =π (f (x)) dx = π = π,
1 1 x2
as we saw recently. So you can easily fill up Gabriel’s horn with water. However in the
homework you will be asked to show that the surface area formula from Calc 1, applied
here, gives that Gabriel’s horn has infinite surface area. So you can fill it up, but you could
never paint it! Interesting... .
.
Chapter 11. Infinite Series.
First we recall ‘sigma-notation’. For example,
4
k2 = 22 + 32 + 42 = 4 + 9 + 16 = 29 .
X
k=2
Here ‘ ’ is read as ‘sigma’ but it should be interpreted as ‘sum of’.
P
Example. Write in sigma-notation: e + 2e2 + 3e3 + 4e4 + 5e5 .

Solution. The kth term here is kek . So the series is 5k=1 kek .
P
Example. Write in sigma-notation: 43 + 54 + 87 + 89 + 12

11 .
Solution. This is almost identical to a sequence example in the previous chapter; the nth term
n+1 n+1 (−1)k+1
is an = 2n+1+(−1) = 1 + (−1)
P5
2n+1 2n+1 . So the series is k=1 (1 + 2k+1 ).
In most of the rest of this course we study ‘infinite series’. These are expressions of the form
∞
X
ak = am + am+1 + am+2 + · · · (∗)
k=m
Let us call this expression (*).

What does expression (*) mean? In fact we shall see shortly that the expression means two
things.
Usually m = 0 or 1, that is, (*) usually is
a0 + a1 + a2 + · · ·
or
a1 + a2 + a3 + · · · .
We call the number ak the kth term in the series. Sometimes we will be sloppy and write k ak
P
when we mean (*).

The most important question about an infinite series, just as for an infinite sequence, is 1) does
the series converge? and 2) if it converges, what is its sum? We will explain these in a minute.
Example 1. Write in sigma notation:

1 1 1
1+ + + + ··· .
2 4 8
Solution. Here the first term in the series is 1, the second term is 12 , the third term is 212 , the fourth
term is 213 , and so on. Clearly the kth term in the series is 2k−1
1
. We may rewrite the series as
∞
X 1
k=1
2k−1
or
∞
X 1
.
k=0
2k

−1 + 1 − 1 + 1 − 1 + · · · .
P∞
Solution. Note that the kth term is (−1)k . The series may be rewritten k=1 (−1)k .
1 1 1 1 1 1 1 1
+ + + + ··· = + + + + ···
2 6 12 20 2·1 3·2 4·3 5·4
1 1 1
Solution. Here the first term in the series is 2·1 , the second term is 3·2 , the third term is 4·3 , and
1
so on. Clearly the kth term in the series is (k+1)k . We may rewrite the series as
∞
X 1
.
k=1
(k + 1)k
We still have not said what expressions like (*) or those in Examples 1, 2, and 3 mean. In fact
an expression like (*), or those in Examples 1, 2, and 3, has two meanings:
Meaning # 1: A ‘formal sum’. That is, it is a way to indicate that we are thinking about
adding up all these numbers in the expression (*), in the order given. It does not mean that these
numbers do add up.
Before we go to Meaning # 2, let me say how you ‘add up all the numbers in an infinite series’.
To do this, we define the nth partial sum sn to be the sum of the first n terms in the series. In this
way we get a sequence
s1 , s2 , s3 , · · ·
called the sequence of partial sums. We say the original series converges if the sequence {sn }
converges. If it does not converge then we say it diverges. We call lim sn the sum of the series if
n→∞
this limit exists.
ak = lim sn if this limit exists.
P
Meaning # 2: k n→∞
Back to Example 2 above: The first meaning of −1 + 1 − 1 + 1 − 1 + · · · (or ∞ k

k=1 (−1) ) is as a
P
‘formal sum’. That is, we we want to think about adding up all the numbers in the expression in
the order given. To get to the second meaning, we must compute the sn ’s. Clearly in this example,
s1 = −1, s2 = −1 + 1 = 0, s3 = −1 + 1 − 1 = −1, s4 = −1 + 1 − 1 + 1 = 0, and so on. In fact the
sequence s1 , s2 , s3 , · · · is the sequence
−1, 0, −1, 0, −1, 0, −1, · · ·
Does this sequence converge? No. So the series diverges. It does not have ‘Meaning # 2’, and it
has no sum.
∞
Back to Example 1 above: The first meaning of 1 + 21 + 14 + 81 + · · · = 1
k=1 2k−1 , is ‘a formal
P
sum’. That is, it is a way to indicate that we would like to add up all these numbers. To actually
add them up, we need to look at the sequence sn of partial sums. In this example, s1 = 1, and s2
is the sum of the first two terms, that is s2 = 1 12 . Also s3 is the sum of the first three terms:
1 1 3
s3 = 1 + + =1
2 4 4
Similarly
1 1 1 4+2+1 7
s4 = 1 ++ + =1 =1 .
2 4 8 8 8
Now you can spot the pattern. The sequence s1 , s2 , s3 , · · · is the sequence
1 3 7 15
1, 1 , 1 , 1 , 1 , · · ·
2 4 8 16
which has limit 2. Thus the original series ∞ 1
P
k=1 2k−1 converges, and its sum is 2. Thus we have
Meaning # 2:
1 1 1
1 + + + + ··· = 2
2 4 8
or
∞
X 1
= 2.
k=1
2k−1
Example 4: Consider the series 1 + 1 + 1 + · · · . Here the nth partial sum sn , namely the sum of
the first n terms, is n. Since lim sn = lim n = ∞ in this case, the series diverges. But its sum,
n→∞ n→∞
namely lim sn , is ∞. So this series does have a Meaning # 2, namely its sum 1+ 1+ 1+ · · · = +∞.
n→∞
∞
Back to Example 3 above: The first meaning of 12 + 61 + 12 1
+ ··· = 1
k=1 (k+1)k , is ‘a formal
P
sum’. That is, it is a way to indicate that we would like to add up all these numbers. To actually
add them up, we need to look at the sequence sn . In this example, s1 = 21 , and s2 is the sum of
the first two terms, that is
1 1 3+1 2
s2 = + = = .
2 6 6 3
Similarly
1 1 1 6+2+1 3
s3 = + + = = .
2 6 12 12 4
Similarly, s4 = 54 , and now you can spot the pattern, it seems that sn = n+1
n
. This can be proved
1
using the method of ‘telescoping series’: to use this method we consider the kth term (k+1)k and
use partial fractions. The partial fraction technique asks you to write
1 A B
= +
(x + 1)x x+1 x
Multiplying through by (x + 1)x gives 1 = Ax + B(x + 1), so that B = 1 and A = −1. Thus
1 1 1
= − .
(k + 1)k k k+1
Therefore setting k = 1, k = 2, k = 3, · · · , k = n in the last equation gives that
1 1 1 1 1 1 1 1 1 1 1 1

sn = + + +···+ = − + − + − +···+ − .
2·1 3·2 4·3 (n + 1)n 1 2 2 3 3 4 n n+1
This is called ‘telescoping’, because most of the last expression cancels, and the expression cancels
to yield
1 n
sn = 1 − =
n+1 n+1
as claimed. Thus the sequence s1 , s2 , s3 , · · · is the sequence
1 2 3 4
, , , , ···
2 3 4 5
which has limit 1. Thus the original series ∞ 1
k=1 (k+1)k converges, and its sum is 1. Thus we have
P
Meaning # 2:
1 1 1
+ + + ··· = 1
2 6 12
or
∞
X 1
= 1.
k=1
(k + 1)k
• In a sum like ∞ 1
P
k=1 (k+1)k , the k is a ‘dummy index’. That is, it is only used internally
inside the sum, and we can feel free to change its name, to ∞ 1
j=1 (j+1)j , for example,
P
• In a series ∞ k=m ak let us call m the ‘starting index’. Thus for example, the starting index
P
of ∞ k−1
k=2 k 2 is 2. Any series can be ‘renumbered’ so that its starting index is 0. That
P
P∞ P∞
is, any infinite series may be rewritten as k=0 ak . For example, k=m ak , which is the
same as am + am+1 + am+2 + P· · · , can be relabelled by letting j = k − m, or equivalently
k = j + m. Then ∞ a = ∞
j=0 aj+m .
P
k=m k
Example: Rewrite ∞ k−1 ∞

P P
k=2 k 2 as a series k=0 ak .
Solution. Letting j = k − 2, so that k = j + 2, the sum becomes
∞ ∞
X j+2−1 X j+1
=
j=0
(j + 2)2 j=0
(j + 2)2
P∞ k+1
Of course j is ‘dummy’ so we can rewrite this as k=0 (k+2)2 .
There is no reason of course why we chose 0 for the starting index. One can make all series begin
with the starting index 1 if you wanted to, by a similar trick. However it is convenient to fix one
starting index, so it may as well be 0. Many of the following results are therfore phrased in terms
of series ∞ k=0 k .
a
P
• Geometric series: This is a series of form c + cx + cx2 + cx3 + · · · , or ∞ k

k=0 cx , for
P
constants c and x. We call x the ‘constant ratio’ of the geometric series. Note that if you
divide any term in the series by the previous term, you get x. We assume c 6= 0, otherwise
this is the trivial series with sum 0.
The MAIN FACT about geometric series, is that such a series converges if and only if
c
|x| < 1, and in that case its sum is 1−x .
[Proof: If c = 0 then the series is 0 + 0 + 0 + · · · and the result is obvious. So we can

assume that c = 6 0. The sum of the first n terms is
sn = c + cx + cx2 + · · · + cxn−1 = c(1 + x + x2 + · · · + xn−1 ) .
There is a well known result in algebra which says that
(1 + x + x2 + · · · + xn−1 )(1 − x) = 1 − xn
(to prove it multiply out the parentheses and cancel). Thus if x 6= 1 then 1 + x + x2 + · · · +
n
xn−1 = 1−x
1−x , so that
1 − xn
sn = c + cx + cx2 + · · · + cxn−1 = c .
1−x
(It is worth memorizing the last formula, the sum of n terms of a geometric series.) The
only thing that depends on n on the right hand side here is the xn , which we saw in 10.4
converges to 0 if |x| < 1, and diverges otherwise. If x = 1 then sn = c + c + · · · + c (n times)
1
which equals nc. Thus lim sn = c 1−x if |x| < 1. If |x| ≥ 1 then {sn } diverges, so that the
n→∞
original series diverges.]
2 2 2
• Example. Is the series 6 + 2 + 3 + 9 + 27 + · · · a geometric series? What is its sum?
Solution: Note that any one term, divided by the previous term, is 31 . So it is a geometric
series, with constant ratio x = 13 and first term c = 6, so that its sum is
6 3
1 = 6· = 9 .
1− 3
2
P∞ 3
• Example. Find the sum of the series k=1 (1.2)k if it converges.
1 3
Solution: This is a geometric series, with constant ratio x = 1.2 and first term c = 1.2 , so
that its sum is
3
1.2 3 3
1 = 1.2 − 1 = 0.2 = 15.
1 − 1.2
• Example: Recurring decimals. Express the number 5.1233333 · · · (recurring) as an

improper fraction.
Solution: 5.1233333 · · · equals
1 2 3 3 3 512 3 3 3
5+ + + + + + ··· = + + + + ··· .
10 100 1000 10000 100000 100 1000 10000 100000
3 1
Since the last part of this is a geometric series with first term 1000 and constant ratio 10 ,
the last sum equals
!
512 3 1 512 3 10 512 1 1537
+ 1 = + = + = .
100 1000 1 − 10 100 1000 9 100 300 300
If ∞
• FACT:P k=0 ak and
∞
k=0 bk both converge, and if Pc is a constant, then:
P P
∞ P∞ ∞
• Pk=0
(ak + b k ) converges, with sum Pk=0
a k + Pk=0
bk ;
∞ ∞ ∞
• Pk=0
(ak − b k ) converges, with sum
P∞ k=0 k
a − k=0 bk ;
∞
• k=0 (cak ) converges, with sum c k=0 ak .
[Proof: We just prove the first one, the others are quite similar. The nth partial sum of
P∞ Pn−1 Pn−1 Pn−1
k=0 (ak + bk ) is k=0 (ak + bk ) = k=0 ak + k=0 bkP
. By a factPabout sums of
sequences from a few weeks ago, this converges, as n → ∞, to ∞ k=0 ak +
∞
k=0 bk .]
(−1)k
P∞
2
• Example: Does k=0 3k
+ 2k
converge? If so, what is its sum?
Solution. It converges. To see this notice that by the previous item we can write
∞ ∞ ∞ ∞ ∞
!
2 (−1)k 2 (−1)k 2 −1 k
X X X X X
+ = + = +
k=0
3k 2k k=0
3k k=0
2k k=0
3k k=0
2
which is a sum of two geometric series. By the rule for geometric series on the previous
page, the series do converge and our sum is
2 1 2
1 + = 3 .
1− 3 1− −1 3
2
• For any positive integer m we can write ∞ a = (a0 + a1 + · · · + am−1 ) + ∞ k=m ak .

P P
P∞ P∞ k
k=0
Indeed k=0 akPconverges if and only if k=m ak converges. P [This is because the nth
partial sum of the ∞ k=0 ak series, and the nth partial sum of the
∞
k=m ak series differ by
a fixed constant.] If these series converge, then their sum also obeys the rule:
∞
X ∞
X
ak = (a0 + a1 + · · · + am−1 ) + ak .
k=0 k=m
P∞ 1 π2 P∞ 1
• Example. You are told that k=1 k 2 = 6 . Find k=3 k 2 .
P∞ 1 1 P∞ 1 pi2 P∞ 1 pi2
Solution. k=1 k 2 = 1+ 2 + k=3 k 2 = 6 . So k=3 k 2 = 6 − 1 − 12 .
• From the last fact it follows that the ‘first few terms’ of a series, do not affect whether the
series converges or not. It will affect the sum though.
• The Divergence Test: If lim ak 6= 0 then the series ak diverges.

P
k
k→∞
A matching statement: If ak converges, then lim ak = 0.
P
k
k→∞
[Beware: If lim ak = 0 we cannot conclude that ak converges.
P
k
k→∞
[Proof: Suppose that ∞ k=0 ak = s. If sn is the nth partial sum then sn → s as n → ∞.
P
Clearly sn+1 → s too, as n → ∞. Thus sn+1 − sn = an → s − s = 0.]

P∞ k
• Example: Determine whether the series k=1 k+1 converges or diverges.
k
Solution: It diverges by the Divergence Test, since lim = 1 6= 0.
k→∞ k+1
P∞ k
• Example: Determine whether the series k=1 (−1)k ln(2k) converges or diverges.
k
Solution: It diverges by the Divergence Test, since lim =∞=
6 0. (One can use
k→∞ ln(2k)
x
LHospital to see that lim = ∞).
x→∞ ln(2x)
P 3k −2
• Example: Determine whether the series k 3k
converges or diverges.
3k −2
Solution: It diverges by the Divergence Test, since lim 3k
= 1 6= 0.
k→∞
11.2-3. Nonnegative Series, and tests for series convergence.

P
• A series k ak is called a nonnegative series if all the terms ak are ≥ 0.
• For a nonnegative series, the sequence {sn } of the partial sums is a nondecreasing (or
increasing) sequence.
[Proof: If sn = a0 + a1 + · · · + an−1 say, then sn+1 = a1 + a2 + · · · + an−1 + an , so that
sn+1 − sn = an ≥ 0. ]
• Therefore, by a fact we saw in Chapter 10 for a nonnegative sequence, the sum of the series
equals the least upper bound of the sequence {sn } of partial sums. Thus the sum of the
series always exists, but may be ∞. More importantly, the series converges if and only
if the {sn } sequence is bounded above. The latter happens if and only if the sum of the
series is finite. Thus to indicate that a nonnegative series converges we often simply write
k k < ∞.
a
P
• Example: The HARMONIC SERIES is the important series
∞
1 1 1 X 1
1 + + + + ··· = .
2 3 4 k=1
k
This is a nonnegative series, so to see if it converges we need only check if the sequence {sn }
is bounded above, where sn = 1 + 21 + 13 + · · · n1 . A trick to do this is to look at 1n+1 x1 dx,
R
interpreted as the shaded area in the graph below [Picture drawn in class]. This shaded
area is less than the area of the n rectangles shown. Hence
1 1 1
Z n+1 1
1 + 1 · + 1 · + ···1 · ≥ dx.
2 3 n 1 x
So sn ≥ ln(n + 1) − ln(1) → ∞ as n → ∞.
Thus the harmonic series diverges; it has sum +∞.
• The trick used in the previous example can be used in the same way to prove:
The Integral Test: If f (x) isPa continuous decreasing positive function
R ∞ defined on [1, ∞)
∞
[Picture drawn in class], then k=1 f (k) converges if and only if 1 f (x) dx converges
(i.e. is finite).
• To redo the previous example using this test, let f (x) = x1 , which is certainly continuous,
decreasing, and positive on [1, ∞). Also
Z ∞ 1
Z c 1
dx = lim dx = lim (ln c − ln 1) = +∞
1 x c→∞ 1 x c→∞
P∞ 1 P∞
So k=1 k = k=1 f (k) does not converge.
P∞ 1
• p-series. An almost identical argument shows that k=1 k p converges if and only if p > 1.
These are called ‘p-series’.
P∞ 1 P∞ 1
• Example. Determine whether the series k=1 √
4 and 3 converge
k=1 k √ or diverge.
k k
Solution: Use the p-series test with p = 1/4 and 4/3. So the first series diverges, the
second converges.
P∞ 1
• Example. Determine whether the series k=3 k(ln k)2 converges or diverges.
Solution: First note that it does’nt matter that the series starts with k = 3, since the
first few terms of a series do not affect whether the series converges or not. Next let
f (x) = x(ln1x)2 , and apply the Integral Test. Note that
Z ∞ 1
Z c 1
dx = lim dx .
3 x(ln x)2 c→∞ 3 x(ln x)2
dx
Substituting u = ln x we have du = x so that
Z c 1
Z ln c 1 1 1 1
dx = du = −u−1 |ln c
ln 3 = − →
3 x(ln x)2 ln 3 u2 ln 3 ln c ln 3
as c → ∞. So both the integral and the series converge.
P∞
• Example. Determine whether the series √1 converges or diverges.
k=3 k ln k
1
Solution: Similar to the last, let f (x) = 1 , and apply the Integral Test. Note that
x(ln x) 2
Z ∞ 1
Z c 1
1 dx = lim 1 dx .
c→∞
3 x(ln x) 2 3 x(ln x) 2
dx
Substituting u = ln x we have du = x so that
Z c 1
Z ln c 1 1 √ √
1 dx = 1 du = 2u 2 |ln c
ln 3 = 2( ln c − ln 3) → ∞
3 x(ln x) 2 ln 3 u 2
as c → ∞. So both the integral and the series diverge.
• BasicPComparison Test: P Suppose that 0 ≤ ak ≤ bk for all k.

1) If Pk bk converges, thenP k ak converges.
2) If k ak diverges, then k bk diverges.
ak P P
• Limit Comparison Test: If 0 < lim bk < ∞, then k ak converges if and only if k bk
k→∞
converges.
• Asked whether a nonnegative series k ak converges, one may first ask: does ak → 0? If
P
not, the answer is ‘no’, by the divergence test. If yes, one may ask how fast does ak → 0? If
very fast, like faster than the terms in a series that you know converges, then it converges
too by something like the Basic Comparison Test.
1
• 11.3. Root Test: If ak is a nonnegative series with lim (ak ) k = r. If 0 ≤ r < 1 then
P
k
k→∞
ak converges. If 1 < r ≤ ∞ then ak diverges.
P P
k k
ak+1
• 11.3. Ratio Test: If ak is a nonnegative series with lim = λ. If 0 ≤ λ < 1 then
P
k ak
k→∞
ak converges. If 1 < λ ≤ ∞ then ak diverges.
P P
k k
P∞
• Examples. Determine whether the following series converge or diverge: (a) √ 1 ;
k=1 k+k 3
P∞ 1√ P∞
√ 1
P∞ k3 P∞ k!
(b) k=1 k+ k ; (c) k=2 k 3 −k
; (d) k=1 2k
; (e) k=3 (2k)! ; (f)
P∞ 1
k=3 (ln k)k .
Solution: (a) The k3 here is much more important that the k here, so P think of the series
1
as being comparable to k 3
. This is a p-series with p = 2 > 1, so k √1 3 converges
P
√
k3 k
by the p-series test above. P Now we can use either the basic comparison test or the limit
comparison test to see that ∞ k=1
√ 1
k+k 3
converges. For example, since k3 < k + k3 we have
√ √ 1
k3 < k + k3 , so that √k+k < √1 3 . Since k √1 3 converges, our other series converges
P
3 k k
by the basic comparison test.
(b) We may start similarly to (a), compare with ∞ 1
P
k=1 k which is divergent. Lets apply
1√ 1
the limit comparison test, with bk = k+ k and ak = k . We have
√
ak k+ k 1
= = 1+ √ →1
bk k k
P∞ 1√
as k → ∞. Since this limit is > 0, the limit comparison test tells us that k=1 k+ k
diverges.
√1
P
(c) Similar to (b), but compare with the convergent p-series k (see (a)), using the
k3
limit comparison test. We let ak = √1 and bk = √ 1 , then
k3 k 3 −k
√ s
ak k3 − k k
= √ = 1− →1>0
bk k3 k3
P∞
as k → ∞. Thus the limit comparison test tells us that √ 1 converges.
k=1 k 3 −k
k3
(d) You could use the ratio or the root test here. If we use the ratio test with ak = 2k
,
(k+1)3
then ak+1 = 2k+1
, so that
3 3
ak+1 (k + 1)3 2k 1 k+1 1 1 1

= k
· 3 = = 1+ →
ak 2·2 k 2 k 2 k 2
as k → ∞. Since this limit is less than 1, the ratio test tells us that the series converges.
k! (k+1)!
(e) We use the ratio test with ak = (2k)! . Then ak+1 = (2(k+1))! . But (k + 1)! may be
written as (k+1)k(k−1) · · · 3·2·1 = (k+1)k!, and similarly, (2k+2)! = (2k+2)(2k+1)(2k)!.
Thus
ak+1 (k + 1)k! (2k)! (k + 1) 1
= · = = → 0
ak (2k + 2)(2k + 1)(2k)! k! (2k + 2)(2k + 1) 2(2k + 1)
as k → ∞. Since this limit is less than 1, the ratio test tells us that the series converges.
1
(f) Converges by the root test with ak = (ln1k)k , since (ak ) k = (ln1k) → 0 as k → ∞.
P∞
• Examples. Determine whether the following series converge or diverge: (a) √ k ;
k=1 1+k 2
P∞ P∞ P∞ P∞
(b) √1 ; (c) √ 1 ; (d) 1
; (e) k!
k=10 k−3 k=1 k 3 +1 k=1 3k +2 k=1 2k ; (f)
P∞ sin( k1 ) √
P∞ 1 k 2 +10
P∞ P∞
√ k
k=1
√ ; (g) k=10 3k 2 −4k+5 (h)
k=10 4k 3 −k 2 +7 ; (i) k=1 ; (j)
k k 3 +1
P∞ e2k P∞ 2k+1 k P∞ k2 2k P∞ k!
k=1 k k ; (k) k=1 ( 3k ) (l) k=1 3k ; (m) k=1 2k .
[These are for extra practice. Some were done in class or review session; all worked in Pam B’s
online notes. Items (a)–(d), (g)–(i) can also be answered very quickly by the ‘winning term’ trick.
k2 1
For example, in (h) the ‘winning terms’ in numerator and denominator give 4k 3 = 4k . So our series
behaves like k 4k = 41 k k1 , which diverges (basically is the harmonic series). So the series in
P 1 P
(h) diverges too. That is the trick. To fully justify this though, if pressed for a proof, one would
1
use the limit comparison test, with bn = 4k .
sin( k1 ) 1 an sin( k1 )
In (f), use the limit comparison test, with an = √
k
and bn = √
k k
. Then bn = 1 → 1 as
k
k → ∞. Since bk is a convergent p-series, the series in (f) converges by the limit comparison
P
k
test. ]
Order of tests for nonnegative series: If you dont recognize it as a geometric or p-series, etc, I’d
use the following order: divergence, limit comparison, root, comparison, integral, ratio. Use the
ratio test if you have factorials, the root test if you have powers, limit comparison test if you have
‘winning terms’, integral if terms are decreasing.
If you are not asked for working, then many such problems may be done instantaneously if you
P ak P ak P k!
use the following rules (memorize): k k! , k kk , k k k all converge rapidly for any number
a. This is because the top grows much much more slowly than the bottom. The same argument
applies to series like k lnkkk , etc.
P
11.4. Absolute and conditional convergence
A series ak is called absolutely convergent if |ak | converges.

P P
k k
• Example. 1+ 14 − 91 + 16
1 1
+ 25 1
− 36 1
+ 49 +· · · is absolutely convergent, because 1+ 14 + 91 + 16
1
+· · ·
converges (by the p-series test with p = 2).
• Key Fact in 11.4: Any absolutely convergent series is convergent.
• We shall see that the converse is false, a series may be convergent, but not absolutely
convergent. Such a series is called conditionally convergent.
• (Proof of Key Fact: If k |ak | converges,Pthen so does k 2|ak |. However 0 ≤ ak + |ak | ≤

P P
2|ak |. So by the basic comparison test, k (ak + |ak |) converges. By the ‘difference rule’
(3rd ‘bullet’ in the FACT towards the end of the notes for Section 11.1 only), ak =
P
k
(a + |a |) − |a | converges.)
P P
k k k k k
1 1 1 1 1 1
• Example. Does series 1 + 4 − 9 + 16 + 25 − 36 + 49 + · · · converge or diverge?
Solution. It converges, since as we saw in the previous example, this series is absolutely
convergent. So by the Key Fact above it is convergent.
sin(kπ 2 )
• Example. Does the series converge or diverge?
P
k k2
sin(kπ 2 ) 2
| converges by the comparison test, since | sin(kπ ) 1
P
Solution. The series k | k2 k2 |≤ k2 ,
sin(kπ 2 )
and k k12 is a convergent p-series. So the series converges absolutely. So by
P P
k k2
the Key Fact above it is convergent.
• The Alternating Series Test: Suppose that a0 > a1 > P a2 > · · · , and that limk ak = 0.
Then a0 − a1 + a2 − a3 + · · · (which in sigma notation is k (−1)k ak ) converges.
[Proof: The 2nth partial sum is
s2n = a0 − a1 + a2 − a3 + · · · − a2n−1 = (a0 − a1 ) + (a2 − a3 ) + · · · + (a2n−2 − a2n−1 ).
Each bracketed term is nonnegative, so that s2 , s4 , s6 , · · · is an increasing sequence, so it
has a limit s say. Similarly
s2n+1 = a0 − (a1 − a2 ) − (a3 − a4 ) − · · · − (a2n−1 − a2n )
and each bracketed term is nonnegative, so that s1 , s3 , s5 , · · · is a decreasing sequence, and
so has a limit t say. But s2n+1 − s2n = a2n which has limit 0 as n → ∞. So s = t and this
is a finite number. Thus {sn } converges.]
• Example. Determine whether the series 1− 12 + 13 − 41 +· · · is convergent, converges absolutely,

converges conditionally, or diverges.
Solution: The series 1 − 12 + 31 − 41 + · · · is convergent by the Alternating Series Test.
But it is not absolutely convergent, because 1 + 12 + 31 + · · · is the divergent harmonic series
(which we met close to the start of Section 11.2). So 1 − 12 + 31 − 41 + · · · is conditionally
convergent. We’ll see later that its sum is ln 2.
• Example. Determine whether the following series are convergent, converges absolutely,
P (−1)k P (−1)k √ √
k ( k + 1 − k) .
converges conditionally. (a) , (b) , (c) (−1)
P
√
k k k 2k k
Solution: Both series are convergent by the Alternating Series Test. But (a) is not
absolutely convergent, because k √1k is divergent by the p-series test of Section 11.2. So
P
(a) is conditionally convergent. Series (b) is a convergent geometric series, and so is k 21k ,
P
so series (b) is absolutely convergent and hence not conditionally convergent. Series (c)
is not absolutely convergent, conditionally convergent, convergent, and not divergent (this
example will probably be done in the review).
• A much more difficult fact to prove is that any ‘rearrangement’ of an absolutely convergent
series is convergent and has the same sum.
• Example. 1 − 19 + 41 − 36
1 1
+ 16 1
− 81 1
+ 25 − · · · converges, since it is a ‘rearrangement’ of the
absolutely convergent series considered at the beginning of this section.
xk
• Example1. For any x show that lim = 0.
k→∞ k!
P xk P |x|k
Solution: Consider the series k k! . This series is absolutely convergent, since k k! is
|x|k
convergent, as one can check using the ratio test: for if ak = k! then as k → ∞
ak+1 |x|k+1 k! |x|
= · k = → 0.
ak (k + 1)! |x| (k + 1)
P xk xk
Thus k k! is convergent, so by the line after the Divergence Test, lim = 0.
k→∞ k!
• If ∞ k
k=1 (−1) ak is a convergent alternating series as in the ‘alternating series test’ above,
P
with sum s, and if sn = nk=1 (−1)k ak , then |sn − s| ≤ |an+1 |. Note that |sn − s| is the
P
error in approximating s by the sum of the first n terms in the series.

P∞ 1
• Example. Approximate the sum of k=1 (−1)k k! by its first six terms, and estimate the
error in your approximation.
P6 1
Solution. k=1(−1)k k! ≈ 0.63194 (calculator). The error in this approximation is less
1
than |a7 | = 7! = 0.00019 (calculator).
P∞ (−1)k+1
• Example. Approximate k=1 k4
with error less than 0.001.
1
Solution: We have by the last bullet that |sn − s| ≤ |an+1 | = (n+1)4 , and this will be less
1
than 0.001 if (n+1)4 < 0.001, or equivalently if (n + 1)4 > 1000. This will be true if n = 5.
(−1) k (−1)k+1
So an approximation to ∞
P5
with error less than 0.001 will be s5 = =
P
k=1 k4 k=1 k4
1 1 1 1
1 − 24 + 34 − 44 − 54 , which is 0.94754 (calculator).
11.5. Taylor Series.

• We are now ready to finish the Calculus 2 syllabus with a discussion of Taylor and Power
series. Up until now we have pretty much proved everything in these notes for Chapter 11
(although you are not expected to read most of these proofs). However from now on the
proofs become too lengthy to include.
1We did this example another way towards the end of Chapter 10.
• The nth Taylor polynomial of a function f (x) is defined to be
f ′′ (0) 2 f ′′′ (0) 3 f (n) (0) n
Pn (x) = f (0) + f ′ (0) x + x + x + ··· + x
2! 3! n!
(k)
which in sigma notation is nk=0 f k!(0) xk .
P
The Taylor series or MacLaurin series of a function f (x) is defined to be

f ′′ (0) 2 f ′′′ (0) 3
f (0) + f ′ (0) x + x + x + ···
2! 3!
P∞ f (k) (0) k
which in sigma notation is k=0 k! x .
• Example. Find the 7th Taylor polynomial of sin x. Also, find the Taylor series of sin x.
Solution: If f (x) = sin(x) then f ′ (x) = cos x, f ′′ (x) = − sin(x), f ′′′ (x) = − cos x, and then
it starts repeating, f (4) (x) = f (x) = sin(x), f (5) (x) = f ′ (x) = cos x, and so on. Thus we
have f (0) = 0 = f (4) (0), f ′ (0) = 1 = f (5) (0), f ′′ (0) = 0 = f (6) (0), f ′′′ (0) = −1 = f (7) (0),
and so on. Thus
−1 3 1 −1 7 x3 x5 x7
P7 (x) = 0 + 1 · x + 0x2 + x + 0x4 + x5 + 0x6 + x = x− + −
3! 5! 7! 3! 5! 7!
and the Taylor series of sin x is
x3 x5 x7
x− + − + ··· .
3! 5! 7!
• Example. Find the Taylor polynomial P4 (x) for ex . Also, find the Taylor series of ex .
Solution: If f (x) = ex then ex = f ′ (x) = f ′′ (x) = f ′′′ (x) = · · · . Thus we have 1 = f (0) =
f ′ (0) = f ′′ (0) = f ′′′ (0) = f (4) (0), and so on. Thus
1 1 1 1 1 3 1
P4 (x) = 1 + x + x2 + x3 + x4 = 1 + x + x2 + x + x4 .
2 3! 4! 2 6 24
The Taylor series of ex is
∞
1 1 xk
1 + x + x2 + x3 + · · · =
X
.
2 3! k=0
k!
• Example. Find the nth Taylor polynomial, and the Taylor series, of ln(1 − x).
Solution: If f (x) = ln(1 − x) then f ′ (x) = −(1 − x)−1 , and
f ′′ (x) = −(−1) · (1 − x)−2 · (−1) = −(1 − x)−2
Similarly, f ′′′ (x) = −(−2)(1 − x)−3 · (−1) = −2(1 − x)−3 , and
f (4) (x) = −(−3) · 2(1 − x)−4 · (−1) = −3 · 2(1 − x)−4
(k−1)!
and so on. In general, the pattern is f (k) (x) = − (1−x) ′
k . Also f (0) = ln(1) = 0, f (0) =
−1, f ′′ (0) = −1, f ′′′ (0) = −2, f (4) (0) = −3 · 2, and in general f (k) (0) = −(k − 1)!. Therefore
f (k) (0) (k − 1)! 1
=− =−
k! k! k
for k = 1, 2, 3, · · · . So the nth Taylor polynomial of ln(1 − x) is
x2 x3 xn
Pn (x) = −x − − − ··· − .
2 3 n
x2 x3 P∞ xk
The Taylor series of ln(1 − x) is −x − 2 − 3 − · · · In sigma notation this is − k=1 k .
• We define the Taylor remainder to be

Rn (x) = f (x) − Pn (x).
This is the error in approximating f (x) by Pn (x). Note
f (x) = Pn (x) + Rn (x).
• Example. Find the Taylor remainder R5 (x) for f (x) = sin x.

x3 x5
Solution: R5 (x) = sin x − P5 (x) = sin x − x + 3! − 5! , by the example above.
• Taylor’s Theorem. If f (n+1) is continuous on an open interval I containing 0, then for

every x ∈ I we have
f ′′ (0) 2 f ′′′ (0) 3 f (n) (0) n
f (x) = f (0) + f ′ (0)x + x + x + ··· + x + Rn (x)
2! 3! n!
1 Rx
where Rn (x) = n! 0 f (n+1) (t)(x − t)n dt. There exists a number c between 0 and x such
that
f (n+1) (c) n+1
Rn (x) = x .
(n + 1)!
This is the Lagrange formula, or the Lagrange form of the remainder. Thus if |f (n+1) (t)| ≤
M for all t between 0 and x then
M |x|n+1
|Rn (x)| ≤ .
(n + 1)!
We can take M to be the maximum of |f (n+1) (t)| on the interval [0, x].
• Note that the Mean Value Theorem from Calculus I is the case n = 0 of Taylor’s Theorem.
• Example. Assume that f is a function such that |f (4) (x)| ≤ 2 for all x. Find the maximal
possible error if P3 ( 12 ) is used to approximate f ( 21 ).
Solution: We may take M = 2. The error in using P3 ( 21 ) to approximate f ( 12 ) is R3 ( 21 ), and
have
1 M | 21 |4 1 1
|R3 ( )| ≤ ≤ 3 = = 0.0052.
2 4! 2 · 4! 192
So the maximal possible error is less than 0.0052.
• Example. Assume that f is a function such that |f (4) (x)| ≤ 2 for all x. Find a small n such
that the maximal possible error is smaller than 0.001, if P3 ( 12 ) is used to approximate f ( 12 ).
Solution: Just as in the previous example, M = 2. The error in using Pn ( 12 ) to approximate
f ( 12 ) is Rn ( 12 ), and have
1 M | 12 |n+1 1
|Rn ( )| ≤ = n
2 (n + 1)! 2 (n + 1)!
1
If we want the error smaller than 0.001 we can solve 2n ·(n+1)! < 0.001. That is, 1000 <
n 4
2 · (n + 1)!. Clearly n = 4 will do that job: 1000 < 2 5! = 16 · 120. So if n = 4 then the
maximal possible error (if P3 ( 12 ) is used to approximate f ( 12 )) is smaller than 0.001.
• Example. Find the Lagrange form of the Taylor remainder Rn for the function f (x) = e2x
and n = 3.
Solution: We have f ′ (x) = 2e2x , f ′′ (x) = 22 e2x , f ′′′ (x) = 23 e2x , f 4 (x) = 24 e2x . So the
Lagrange form is
f (4) (c) 4 24 e2c 4 2e2c x4
R3 (x) = x = x = .
4! 4! 3
• KEY POINT: Up until now, we have not considered at all the question of whether the
Taylor series of f (x) converges. Clearly it does converge when x = 0 to f (0) though. The
partial sums of the Taylor series of f (x) are just the Pn (x), so it follows that the Taylor series
converges to f (x) if and only if Pn (x) → f (x), or equivalently, if and only if lim Rn (x) = 0.
n→∞
To check if lim Rn (x) = 0, the last fact in Taylor’s theorem is very useful, as we shall
n→∞
now see in many examples:
3 5 7
• Example 1. Show that the series x − x3! + x5! − x7! + · · · converges to sin x for every number
x. Thus for every number x,
x3 x5 x7
sin x = x − + − + ···
3! 5! 7!
Solution: The series in this example is the Taylor series of f (x) = sin x, as we saw on the
previous page. To show that the series converges to sin x for all x, by the KEY POINT
above, we need to show that lim Rn (x) = 0. Note that |f (n+1) (x)| = | sin x| or | cos x|, so
n→∞
that |f (n+1) (x)| ≤ 1 for every number x. So we can take M = 1 in the last part of Taylors
theorem, and
|x|n+1
|Rn (x)| ≤
(n + 1)!
But the right hand side here converges to 0 as we saw close to the top of page 9. So by the
squeezing or pinching rule lim Rn (x) = 0.
n→∞
• Example 2. An exactly similar argument shows that for every x,

x2 x4 x6
cos x = 1 − + − + ···
2! 4! 6!
∞
x2 x3 xk
ex = 1 + x +
X
+ + ··· =
2! 3! k=0
k!
• Example 3. Using the example a couple of pages back of the Taylor series of ln(1 − x), show
(a) that 1 − 12 + 13 − 14 + · · · has sum equal to ln 2.
(b) Approximate ln 2 with error less than 0.05, using a ‘Taylor approximation’.
(c) Find a value of n so that Pn (x) approximates ln 2 with error less than 0.001.
Solution: (a): Let f (x) = ln(1 − x). We saw in an example a couple of pages back that the
2 3
Taylor series of ln(1 − x) is −x − x2 − x3 − · · · . Putting x = −1 we see that the Taylor
series of ln(1 − x) is 1 − 21 + 13 − 14 + · · · when x = −1. And f (−1) = ln 2. But this is
not a proof that 1 − 12 + 31 − 41 + · · · = ln 2 yet. What we need to show is that the Taylor
series converges to f (x) when x = −1, but by the KEY POINT above, this is the same as
showing that Rn (−1) → 0 as n → ∞. So let us show the latter.
We also saw in that example a couple of pages back that f (n+1) (t) = − (1−t)n!(n+1) , so if
−1 ≤ t ≤ 0 then 1 − t ≥ 1, so that |1 − t|n+1 ≥ 1, so that we have
n!
|f (n+1) (t)| = ≤ n!.
|1 − t|n+1
So we can take M = n! and x = −1 in the last line of Taylors theorem and get
n! 1
|Rn (−1)| ≤ = →0
(n + 1)! n+1
as n → ∞. Thus by the KEY POINT above, the Taylor series of ln(1 − x) when x = −1,
which is 1 − 12 + 13 − 41 + · · · , converges to f (−1) = ln(2).
To get (b), we want the Taylor remainder |Rn (−1)|, which is the error in approximating
1 1
f (−1) by Pn (−1), to be less that 0.05. We just saw that |Rn (−1)| ≤ n+1 . So if n+1 < 0.05
1 1 1 1 1
we will be done. But n+1 < 0.05 if n ≥ 20. Thus P20 (−1), which equals 1− 2 + 3 − 4 +· · ·− 20 ,
is a Taylor approximation to ln 2 with error less than 0.05.
1
Item (d) is just like (c), we want |Rn (−1)| < 0.001. In (a) we saw that |Rn (−1)| ≤ n+1 .
1 1
So if n+1 < 0.001 we will be done. But n+1 < 0.001 if n + 1 > 1000. So choose n = 1000.
π π
• Example. Estimate the error if P4 ( 10 ) is used to approximate sin( 10 ). If you do not have
a calculator do not simplify your estimate too much, but it should be a number (that is, it
should have no variables in it).
π
Solution: We want to estimate R4 ( 10 ), when f (x) = sin x. Since |f (5) | = | cos x| ≤ 1, in the
estimate for Rn (x) in the last line of Taylor’s Theorem, we can take M = 1, and so that
estimate becomes:
π 1 π 5 π5
|R4 ( )| ≤ ( ) = .
10 5! 10 120 · 105
π5
Final answer: 120·105.
• Example. Let Pn be the nth Taylor Polynomial of the function f (x). Assume that f is
a function such that |f (n) (x)| ≤ 1 for all n and x (the sine and cosine functions have this
property.) Find the least integer n for which Pn (0.5) approximates f (0.5) to within 0.001.
Solution: In the estimate for Rn (x) in the last line of Taylor’s Theorem, we can take M = 1,
and so that estimate becomes:
1 1
|Rn (0.5)| ≤ (0.5)n+1 = n+1 ,
(n + 1)! 2 (n + 1)!
and this is < 0.001 if 2n+1 (n + 1)! > 1000. This is true if n = 4 (since 25 5! = 3840).
‘Answer’: n = 4.
11.6. The Taylor series of a function f (x) about a number a, is the series
f ′′ (a) f ′′′ (a)
f (a) + f ′ (a) (x − a) + (x − a)2 + (x − a)3 + · · ·
2! 3!
P∞ f (k) (a)
which in sigma notation is k=0 k!
(x − a)k .
• The nth Taylor polynomial of a function f (x) about a number a, is defined to be
f ′′ (a) f ′′′ (a) f (n) (a)
Pn (x) = f (a) + f ′ (a) (x − a) + (x − a)2 + (x − a)3 + · · · + (x − a)n
2! 3! n!
and the Taylor remainder Rn (x) = f (x) − Pn (x).
• Example. Suppose that g is a function which has continuous derivatives, and that
g(2) = 3, g ′(2) = −4, g ′′ (2) = 7, g ′′′ (2) = −5. Find the Taylor polynomial of degree 3
for g centered at x = 2.
Solution: P3 (x) = 3 − 4(x − 2) + 72 (x − 2)2 − 65 (x − 2)3 .
• Example. Find the Taylor series of f (x) = x2 + 3x − 1 about x = 1.

Solution: f (1) = 3, f ′(x) = 2x + 3, f ′(1) = 5, f ′′ (x) = 2, f ′′′ (x) = 0, and so on. So the
Taylor series about x = 1 is
f ′′ (1)
f (1) + f ′ (1) (x − 1) + (x − 1)2 + 0 + 0 + · · · = 3 + 5(x − 1) + (x − 1)2 .
2!
• Example. Expand f (x) = x2 + 3x − 1 in powers of (x − 1).

Solution: This is just another way to ask for the Taylor series about x = 1. So the
answer is 3 + 5(x − 1) + (x − 1)2 .
• Example. Find the Taylor series of ln x about x = 1. Also find the nth Taylor
polynomial of ln x about x = 1.
Solution: Let f (x) = ln x. Then f ′ (x) = x−1 , f ′′ (x) = −x−2 , f ′′′ (x) = +2x−3 , f (4) (x) =
−3·2x−4 . In general f (k) (x) = (−1)k−1 ·(k −1)!x−k . Thus f (1) = 0, f ′(1) = 1, f ′′ (1) =
−1, f ′′′ (1) = 2, f (4) (1) = −3 · 2, and in general f (k) (1) = (−1)k−1 · (k − 1)!. Thus
f (k) (1)
k!
= (−1)k−1 k1 . So the Taylor series about x = 1 is
(x − 1)2 (x − 1)3 (x − 1)4
(x − 1) − + − +··· .
2 3 4
The nth Taylor polynomial about x = 1 is
(x − 1)2 (x − 1)3 1
Pn (x) = (x − 1) − + − · · · + (−1)n−1 xn .
2 3 n
• Just as in 11.5 we have Taylor’s Theorem: If f (n+1) is continuous on an open

interval I containing a, then for every x ∈ I we have
f ′′ (a) f ′′′ (a) f (n) (a)
f (x) = f (a) + f ′(a) (x − a) + (x − a)2 + (x − a)3 + · · · + (x − a)n + Rn (x)
2! 3! n!
1 Rx (n+1) n
and Rn (x) = n! a
f (t)(x − t) dt, and indeed
f (n+1) (c)
Rn (x) = (x − a)n+1
(n + 1)!
for some number c between a and x. This is the Lagrange formula, or the Lagrange
form of the remainder. Thus if |f (n+1) (t)| ≤ M for all t between a and x then
M|x − a|n+1
|Rn (x)| ≤ .
(n + 1)!
Again we can take M to be the maximum of |f (n+1) (t)| for t between a and x.
• Example. Using the last example, approximate ln(1.1) with error less than 0.001.
Solution: Let a = 1, x = 1.1 in Taylor’s Theorem above. In the previous example we
saw that f (n+1) (t) = (−1)n · n!t−n−1 , so that |f (n+1) (t)| = n!t−n−1 ≤ n!, if 1 ≤ t ≤ 1.1.
From the last line of Taylor’s Theorem,
n! (0.1)n+1 (0.1)n+1
|Rn (1.1)| ≤ = .
(n + 1)! n+1
Notice if n = 2 then |R2 (1.1)| ≤ 0.001
3
< 0.001. So the approximation we want is
the 2nd Taylor polynomial, which by what we did in the previous example will be
2 2
P2 (x) = (x − 1) − (x−1)
2
. Thus ln(1.1) is approximately P2 (1.1) = (0.1) − (0.1)
2
=
0.1 − 0.005 = 0.095.
√
• Example. Find the Taylor polynomial P3 (x) for the function f (x) = x about √ x = 4.
Also find the Lagrange form of the remainder R2 (x) for the function f (x) = x about
x = 4.
1 3 5
Solution: We have f ′ (x) = 12 x− 2 , and f ′′ (x) = − 21 21 x− 2 , and f ′′′ (x) = 38 x− 2 . So
1 3 5
f (4) = 2, f ′(4) = 12 4− 2 = 41 , f ′′ (4) = − 14 4− 2 = − 32
1
, f ′′′ (4) = 83 4− 2 = 8·32
3 3
= 256 . So
1 1 2 1 3
P3 (x) = 2 + 4 (x − 4) − 64 (x − 4) + 512 (x − 4) .
The Lagrange form of the remainder is
3 −2 5 5
f (3) (c) c c2
R2 (x) = (x − 4)3 = 8 (x − 4)3 = (x − 4)3 .
3! 6 16
Here c is a number between 4 and x.
11.7. Power Series.

P∞
• A power series is a series of the form k=0 ck xk , or in longhand,
c0 + c1 x + c2 x + c3 x3 + · · ·
2
where c0 , c1 , c2 , · · · are constants.
• Example 1. 1 + x + x2 + x3 + · · · .
x2 x4 x6
• Example 2. 1 − 2!
+ 4!
− 6!
+··· .
P∞ xk
• Example 3. k=1 k
.
• KEY RESULT IN 11.7: For any power series ∞ k

k=0 ck x , there is a constant R,
P
called the radius of convergence of the power series. We have 0 ≤ R ≤ ∞ and

(a) If R = 0 then the power series only converges when x = 0.
(b) If R = ∞ then the power series converges for every number x.
(c) If 0 < R < ∞ then the power series converges if −R < x < R, and it diverges if
|x| > R.
Thus the set of numbers x for which the power series converges is an interval
centered at the origin on the number line. This interval is called the interval of
convergence. [Picture drawn in class].
• Examples. Find the radius of convergence and the interval of convergence for each of
Examples 1, 2, 3 above.
Solutions: Example 1 above is a geometric series, and from the geometric series
test, we know it converges only for x ∈ (−1, 1). This is its interval of convergence.
Its radius of convergence is therefore 1.
Example 2: we said in 11.5 that this series converges for all x to cos x. So its
interval of convergence is (−∞, ∞), and its radius of convergence is therefore ∞.
Put x = 1 in Example 3, and one gets the divergent harmonic series (see Section
11.2). Put x = −1 in Example 3, and one gets a convergent alternating series (this
was a major example in Section 11.4). Thus the interval of convergence must be
[−1, 1), and the radius of convergence is 1.
• A formula for the radius of convergence R:
|ck |
R = lim
k→∞ |ck+1|
if this limit exists.
• Example 4. Find the radius of convergence and the interval of convergence for the
power series ∞ k
k=1 k x .
P
P∞
Solution: The radius of convergence of the power series k=1 k xk is 1 since
k 1
lim = lim 1 = 1.
k→∞ k+1 k→∞ 1+ k
Since this series diverges when x = 1 or x = −1, but the radius of convergence is 1,
the interval of convergence must be (−1, 1).
• If the limit in the ‘formula for R’ above does not exist, try
1
R = lim 1
k→∞ |ck | k
if this limit exists.

[These two ‘formulae for R’ are easily proved using the ‘ratio test’ and ‘root test’. ]
• The sum function of a power series: Suppose that ∞ k

k=0 ck x is a power series,
P
and suppose that I is its interval of convergence. For x in I, define f (x) to be the sum
of the series; that is f (x) = ∞ k
k=0 ck x (in the sense of Meaning # 2 of the sum of a
P
series (see 2nd page of these typed notes on Chapter 11)). Then f : I → (−∞, ∞) is
a function defined on the interval of convergence. We call f the sum function. It is
actually a continuous function on I, and differentiable in the interior of I as we shall
see in the next section.
• Examples. In Example 1 above, the sum function is, by the geometric series formula,
1
f (x) = 1−x . In Example 2, the sum function is cos x.
k
In Example 3, the sum function f (x) = ∞ x
is defined on the interval of
P
k=1 k
convergence. On the last page we saw that the interval of convergence is [−1, 1). So
P∞ xk
k=1 k is a continuous function on [−1, 1), and it is differentiable on (−1, 1).
In Example 4, the sum function f (x) = ∞ k
k=1 k x is defined on the interval of
P
convergence. What is this interval? Since this series diverges when x = 1 or x = −1,
but the radius of convergence is 1, the interval of convergence must be (−1, 1). So
P∞ k
k=1 k x is a continuous function on (−1, 1).
• The results in this section above have variants for power series ∞
k=0 ck (x − a)
k
P
about a number a. One simply replaces 0 by a, and x by x − a in places, etc. The

radius of convergence R still has the formula R = lim 1 1 if this limit exists (or
k→∞ |ck | k
|ck |
lim |ck+1 |
). The interval of convergence of this power series about x = a will be
k→∞
either [a − R, a + R], (a − R, a + R], [a − R, a + R), or (a − R, a + R) [Picture drawn in
class]. Outside of the interval of convergence it diverges, but inside its sum function
is continuous.
• Example 4. Find the radius of convergence and the interval of convergence for the
power series ∞ k
k=1 k (x − 1) .
P
P∞
Solution: The radius of convergence of the power series k=1 k (x − 1)k is 1 since
k 1
lim = lim 1 = 1.
k→∞ k+1 k→∞ 1+ k
Since this series diverges when x = 0 or x = 2, but the radius of convergence is 1, the
interval of convergence must be (0, 2).
11.8. Differentiation and integration of power series.

• The differentiated power series of a power series c0 +c1 x+c2 x2 +c3 x3 +c4 x4 +· · · ,
is the power series
c1 + 2c2 x + 3c3 x2 + 4c4 x3 + · · ·
• Example. The differentiated power series of the power series 1 + x + x2 + x3 + · · · is

the power series 1 + 2x + 3x2 + 4x3 + · · · .
• KEY RESULTS
P∞
ON DIFFERENTIATING POWER SERIES: Suppose that a power
k
series k=0 ck x has a radius of convergence R > 0. Then:
A) the differentiated power series has the same radius of convergence R.
Let f (x) be the sum function of the power series (defined on the previous page),
and let g(x) be the sum function of the differentiated power series.
B) f (x) is differentiable on (−R, R), and f ′ (x) = g(x) for all x in (−R, R). That
is, if −R < x < R then
∞
′
kck xk−1 .
X
f (x) =
k=1
C) We can iterate this process, and look at the ‘second differentiated P∞
power se-
ries’ (i.e, the differentiated power series of differentiated power series) k=2 k(k −
1)ck xk−2 . By A) and B), this has the same radius of convergence R, and its sum
function equals f ′′ (x) on (−R, R). Thus if −R < x < R then
∞
f ′′ (x) = k(k − 1)ck xk−2 .
X
k=2
Similarly for f ′′′ (x), f (4) (x), and so on.

D) Putting x = 0 in f (x), f ′(x), f ′′ (x), f ′′′ (x) and so on, we find from B) and C) that
(k)
f (0) = c0 , f ′(0) = c1 , f ′′ (0) = 2c2 , and more generally, f (k) (0) = k!ck , or ck = f k!(0) .
P∞
E) Because of D), the Taylor series of f (x) is our original power series k=0 ck xk .
• Part E is saying that the sum function f (x) of a power series on its interval of
convergence, has Taylor series equal to the original power series.
• Example: Find a simple formula for the sum of the series x + 2 x2 + 3 x3 + 4 x4 + · · · .

1
Solution: The geometric series 1 + x + x2 + · · · converges to 1−x for x ∈ (−1, 1).
2 d 1
Differentiating this, we have by B) above that 1+2x+3x +· · · converges to dx ( 1−x )=
−2 2 3 4
(1 − x) , for all x ∈ (−1, 1). Multiplying by x shows that x + 2 x + 3 x + 4 x + · · ·
converges to x (1 − x)−2 for x ∈ (−1, 1). This is the desired sum function.
• Equality of power series. Suppose that ∞ k=0 akP xk and ∞ k

k=0 bk x are two power
P P
P∞ k ∞ k
series whose sums are equal (i.e. k=0 ak x = k=0 bk x ) for all x in an open
interval containing 0. Then a0 = b0 , a1 = b1 , · · · , ak = bk for every k.
[Proof: Let f (x) = ∞ k P∞
k=0 ak x = bk xk for all x in this interval. By D), we
P
k=0
(k)
have ak = f k!(0) = bk for every k.]
• The integrated power series of of a power series c0 +c1 x+c2 x2 +c3 x3 +c4 x4 +· · · ,
is the power series
x2 x3 x4
c0 x + c1 + c2 + c3 +··· .
2 3 4
• RESULT ON INTEGRATING POWER SERIES: Suppose that ∞ k

k=0 ck x is a
P
power series with radius of convergence R > 0. Then the integrated power series has
the same radius of convergence R. If f (x) is the sum functionR of ∞ k
k=0 ck x , and if
P
x
F (x) is the sum function of the integrated power series, then 0 f (t) dt = F (x), for
|x| < R.
x3 5 7
• Example 1. Show that tan−1 x = x − 3
+ x5 − x7 + · · · , for −1 < x < 1.
Solution: Consider the series 1 − x2 + x − x + · · · . This is a geometric series which
4 6
1
converges for −1 < x < 1, with sum 1+x 2 . By the last bullet,
x3 x5 x7 x dt
Z
x− + − +··· = 2
= tan−1 x − tan−1 0 = tan−1 x
3 5 7 0 1+t
x3 x5 x7
• Example 2. Show that tan−1 x = x − 3
+ 5
− 7
+ · · · , for −1 ≤ x ≤ 1.
Solution: Note that the power series here converges when x = 1 and when x = −1,
by the Alternating Series Test (from Section 11.4). So the sum function f (x) of this
power series is continuous on [−1, 1] by the second last bullet or so in Section 11.7
above. Hence
f (1) = lim− f (x) = lim− tan−1 x = tan−1 1.
x→1 x→1
That is, tan−1 1 = 1 − 31 + 15 − 17 + · · · . Similarly f (−1) = tan−1 (−1).
• Example 3. Show that π = 4(1 − 31 + 15 − 17 + · · · ).

Solution: By the previous example π
4
= tan−1 1 = 1 − 31 + 51 − 17 + · · · .
• Example. Note cosh x = 21 (ex + e−x ). But
x −x x2 x3 x2 x3 x2 x4
e +e = (1 + x + + + · · · ) + (1 − x + − + · · · ) = 2(1 + + + · · · ).
2! 3! 2! 3! 2! 4!
x2 x4
Thus cosh x = 21 (ex + e−x ) = 1 + 2!
+ 4!
+ · · · . Similarly,
1 x3 x5
sinh x = (ex − e−x ) = x + + +··· .
2 3! 5!
• Example. Expand x2 cos x3 in powers of x.

Solution: We know that
x2 x4 x6
cos x = 1 − + − +··· .
2! 4! 6!
Thus
(x3 )2 (x3 )4 (x3 )6 x6 x12 x18
cos x3 = 1 − + − +··· = 1− + − +··· .
2! 4! 6! 2! 4! 6!
Thus
x8 x14 x20
x2 cos x3 = x2 − + − +··· .
2! 4! 6!
• Read Abel’s theorem, bottom of page 693 in text.
• Similar results about differentiation and integration hold for power series about a
point x = a. For example: If c0 + c1 (x−a) + c2 (x−a)2 + c3 (x−a)3 + c4 (x−a)4 + · · ·
is a power series with radius of convergence R > 0, then the ‘differentiated power
series’ is the power series
c1 + 2c2 (x − a) + 3c3 (x − a)2 + 4c4 (x − a)3 + · · · ,
and this has the same radius of convergence R. If f (x) is the sum function of the
power series, and if g(x) be the sum function of the differentiated power series, then
f (x) is differentiable on (a − R, a + R), and f ′ (x) = g(x) for all x in (a − R, a + R).
That is, if a − R < x < a + R then
∞
f ′ (x) = kck (x − a)k−1 .
X
k=1
The integrated power series is the power series
(x − a)2 (x − a)3 (x − a)4
c0 (x − a) + c1 + c2 + c3 +··· .
2 3 4
The integrated power series has the same radius of convergence R. If f (x) is the sum
function of ∞ k
k=0 ck x , and if F (x) is the sum function of the integrated power series,
P
Rx
then 0 f (t) dt = F (x), for |x − a| < R.
END OF COURSE

Chapter 10. Sequences, Etc.: Upper Bound

Uploaded by

Copyright:

Available Formats

Chapter 10. Sequences, Etc.: Upper Bound

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Chapter 10. Sequences, Etc.: Upper Bound

Uploaded by

Copyright:

Available Formats

Chapter 10. Sequences, etc.

10.1: Least upper bounds and greatest lower bounds.

• A sequence is a numbered string of objects a1 , a2 , a3 , · · · . We often write the sequence as

• Example 1. List (n2 − 2n + 1)∞ n=1 as a string of numbers.

• Example 2. List (n2 − 2n + 1)∞ 2 4

A sequence is called monotone if it is either increasing or decreasing. The boundedness and

10.3: Limits of sequences.

An unbounded sequence must be divergent.

10.4: Some important limits.

• xn → 0 as n → ∞ if |x| < 1 (this is proved like in the example in Section 10.3 of ( 32 )n ).

10.5–10.6: L’Hospitals rule.

• Example. Find limx→0+ xx .

• Example. Show that limx→∞ (1 + xa )x = ea , and that limx→∞ (1 + x1 )x = e.

10.7: Improper integrals.

• Example. Determine whether or not the integral 1∞ dx

• Example. Determine whether or not the integral 1∞ dx

• Example. Determine whether or not the integral 01 √ dx R

• Example. Determine whether or not the integral 01 ( √1x − (x−1)1 R

both at 0 and at 1. Let us write

• A ”Comparison test”: If 0 ≤ f (x) ≤ g(x) on my interval, and if the integral of g on the

Example. Write in sigma-notation: e + 2e2 + 3e3 + 4e4 + 5e5 .

Example. Write in sigma-notation: 43 + 54 + 87 + 89 + 12

Let us call this expression (*).

when we mean (*).

Example 1. Write in sigma notation:

Example 2. Write in sigma notation:

Back to Example 2 above: The first meaning of −1 + 1 − 1 + 1 − 1 + · · · (or ∞ k

Example: Rewrite ∞ k−1 ∞

• Geometric series: This is a series of form c + cx + cx2 + cx3 + · · · , or ∞ k

[Proof: If c = 0 then the series is 0 + 0 + 0 + · · · and the result is obvious. So we can

sn = c + cx + cx2 + · · · + cxn−1 = c(1 + x + x2 + · · · + xn−1 ) .

There is a well known result in algebra which says that

• Example: Recurring decimals. Express the number 5.1233333 · · · (recurring) as an

• For any positive integer m we can write ∞ a = (a0 + a1 + · · · + am−1 ) + ∞ k=m ak .

• The Divergence Test: If lim ak 6= 0 then the series ak diverges.

Clearly sn+1 → s too, as n → ∞. Thus sn+1 − sn = an → s − s = 0.]

11.2-3. Nonnegative Series, and tests for series convergence.

as c → ∞. So both the integral and the series diverge.

• BasicPComparison Test: P Suppose that 0 ≤ ak ≤ bk for all k.

A series ak is called absolutely convergent if |ak | converges.

• Key Fact in 11.4: Any absolutely convergent series is convergent.

• (Proof of Key Fact: If k |ak | converges,Pthen so does k 2|ak |. However 0 ≤ ak + |ak | ≤

• Example. Determine whether the series 1− 12 + 13 − 41 +· · · is convergent, converges absolutely,

error in approximating s by the sum of the first n terms in the series.

11.5. Taylor Series.

The Taylor series or MacLaurin series of a function f (x) is defined to be

• We define the Taylor remainder to be

• Example. Find the Taylor remainder R5 (x) for f (x) = sin x.

• Taylor’s Theorem. If f (n+1) is continuous on an open interval I containing 0, then for

• Example 2. An exactly similar argument shows that for every x,

• Example. Find the Taylor series of f (x) = x2 + 3x − 1 about x = 1.

• Example. Expand f (x) = x2 + 3x − 1 in powers of (x − 1).

• Just as in 11.5 we have Taylor’s Theorem: If f (n+1) is continuous on an open

11.7. Power Series.

where c0 , c1 , c2 , · · · are constants.

• KEY RESULT IN 11.7: For any power series ∞ k

called the radius of convergence of the power series. We have 0 ≤ R ≤ ∞ and

• A formula for the radius of convergence R:

if this limit exists.