Professional Documents
Culture Documents
2.1
2.1.1
Sequences
Definition of a sequence
1 1 1 1 , , ,........, ,...... 2 3 4 n
0, 1, 0, 2, 0, 3, .
In each of the above we have an example of a sequence. In each of them we have an endless list of numbers and these numbers are listed in order. Let us call these numbers the terms of the sequence. In example (1): The first term is 1, the second term is
1 1 , the third term is and so on. 2 3 1 For any given positive integer n, the nth term is . This is called the n 1 general nth term and we denote the sequence by . n 1 By just putting the values 1, 2, 3 etc for n in , we get that the first term n 1 1 is 1, the second term is , the third term is etc. 2 3
In example (2): The first term is 1, the second term is -1, the third term is 1, the fourth term is -1 and so on. i.e., we know for any positive integer n what the nth term would be. If n is odd it will be 1 and if n is even it will be -1. We can express the nth term as (-1)n +1 since (-1)n +1 = 1 when n is odd and (-1)n +1 = -1 when n is even. So the sequence is ( 1) n +1 .
In example (3): The general nth term is 2n + 1. When n = 1, 2n + 1 = 3 and when n increases by 1, 2n + 1 increases by 2. So we get that, when n = 2, 2n + 1 = 5, when n = 3, 2n + 1 = 7 etc. So, the sequence is 2n + 1 . In example (4): The sequence is
1 2 n 1
In example (5): Although we have not given the general nth term, we know that given any value for n, we can get the value of the nth term. We proceed along 0, 1, 0, 2, 0, 3, and so on until we come to the nth term. Let us get the 16th term in this way: We have, 0, 1, 0, 2, 0, 3, 0, 4, 0, 5, 0, 6, 0, 7, 0, 8. The 16th term is 8. n We see that when n is even, then nth term is and when n is odd, the nth 2 term is 0.
When speaking about a sequence in general, we denote it by a n (or x n , bn etc.) and an denotes the general nth term. We can express the sequence in example 5 by a n where an = when n is odd.
1 . 1 + an n when n is even and an = 0 2
The above equation gives a sequence a n once the value of a1 is given. (i) Let us take it that a1 = 0. Then 1 1 1 1 1 a2 = = = 1, a3 = = , a4 = = 1 + a3 1 + a2 2 1 + a1 1 + 0
1 1+ 1 2
2 etc. 3
(ii)
1 2 1 1 = , a3 = = etc. 1 + a2 3 1 + a1 2
(iii)
5 1 . 2
5 1 5 +1 = and 2 2 1 2 2( 5 1) 5 1 1 = = = . i.e., =. 1+ 5 1 2 1+ 5 +1 1 1 1 1 So, a 2 = = = , a3 = = = etc, 1 + a1 1 + 1 + a2 1 + That is, for any positive integer n, an = . We say here that a n is a constant sequence since all the terms take the
same value. Example (7): x n +1 = x n + x n 1 for all n 2 . The above equation gives a sequence x n once the values of x1 and x2 are given.
2.1.2
Examples: (1)
1 1 1 1 . i.e., 1, , , ,........ . We see that as n grows n 2 3 4 1 approaches the value 0. We say that the indefinitely large, the value of n 1 1 sequence converges to 0 as n tends to infinity and we write lim = 0 . n n n
(2)
n 1 2 3 4 5 . i.e., , , , , , ....... n +1 2 3 4 5 6
n = 1. n n + 1
In examples (1) and (2) we have what we call convergent sequences. (3) Consider the sequence (1) n . i.e., -1, 1, -1, 1, -1, 1, .
There is no number l such that (-1)n approaches the value l as n becomes indefinitely large. For this reason we say that (1) n is a divergent sequence. (4) Consider the sequence 3n 7 . i.e., -4, -1, 2, 5, 8, 11, 14, . The is no number l such that 3n 7 approaches the value l as n becomes indefinitely large. So, 3n 7 is a divergent sequence.
2.1.3. Limits of a sequence Definition 1: Suppose a n is a sequences and l is a real number. We say that, a n converges to l and write lim a n = l if given any real number such that
n
> 0, there is a positive integer n0 such that, whenever n > n0, an l < .
(Note: This is the technical way of saying that an approaches the value l as n becomes indefinitely large). In this case we also say that the sequence a n is convergent. When for a sequence a n there is no number l such that lim a n = l , i.e.,
n
n . n +1
if n + 1 >
. i.e., if n >
1
- 1.
- 1.
(i)
We say that x n
any real number k, there is a positive integer no such that when n > no, xn > k. (Note: This is a technical way of saying that xn grows indefinitely large as n becomes indefinitely large.) (ii) We say that
given any real number k, there is a positive integer no such that when n > no, xn < k. (Note: This is a technical way of saying that xn grows indefinitely 1 1 small as n becomes indefinitely large. (note: is smaller than , 2 3 -100 is smaller than -2, etc.))
Let us apply definition 2(i) to the sequence n 2 . Let k R and n N. n 2 n. Therefore, n2 > k when n > k. We can find no such that no > k. So, for all n, n2 > k when n > no. Therefore, lim n 2 = .
n
2.1.4. Elementary Properties of Limits Suppose a n , bn are sequences and c R . Then, ca n denotes the sequence whose nth term is can, a n + bn denotes the sequence whose nth term is an + bn etc. Like this, given two sequences, we can form other sequences by subtracting, multiplying, dividing. In this spirit, a n denotes the sequence whose nth term is an.
Theorem 1 (Algebra of Limits): Suppose x n , y n are convergent sequences and lim x n = l1 and lim y n = l 2 . Suppose
n n
k R and a n is the constant sequence where for all n N, an = k. Suppose c R. Then: (i) (i.e., lim k = k ) lim a n = k .
n n
lim ( x n + y n ) = l1 + l 2 .
n n
lim
n
x n l1 = yn l2
(i.e., lim(
n
lim x n
lim x n = l1 .
n
1 = 0. n n
Now, by applying the theorem we get 1 1 1 1 lim ( 2 ) = 0, lim ( 3 ) = 0, lim ( 2 3 ) = 0, etc. n n n n n n n Consider for example, lim a n where a n =
n
n 3 3n 2 + 1 . 4n 3 + 5
3 1 + n n 3 and we see that by repeatedly applying the theorem we get, Then, an = 5 4+ 3 n 1 1 = . lim a n = n 4 4 1
The limits in the theorem are finite limits. Definition 1 is about finite limits. definition 2 we have infinite limits. (i.e., lim a n = , lim bn = )
n n
In
What about the properties of infinite limits? For instance, we have that if lim x n = and lim y n = , then lim ( x n + y n ) = .
n n n
This can be coded as + = . We now give some properties of infinite limits in code. Theorem 2: 6
+ =
= ( ) = ( ) ( ) = 1 1 = 0 and =0
+ l = . Suppose l R and l is a constant. Then: (i) If l > 0, l = and l ( ) = (ii) If l < 0, l = and l ( ) =
l = 0 and + l = and
Therefore lim(n + n ) = .
5 3 n
1 1) and n2
1 1) = 1 . Therefore, lim(n 5 + n 3 ) = . 2 n n
n n0 , a n c n bn . Now if lim a n = lim bn = l for some l R , then lim c n = l . Example: 0 sin as 1).
lim 0 = lim
n
1 1 for all positive integers n (i.e., we can take the n0 in the theorem n n
1 1 = 0 . Therefore, lim (sin ) = 0 . n n n n
Then: (i)
(ii)
1 1 1 Example: Consider the sequence a n where for all n, a n = 1 + ( + + .... + ) (i.e., 1! 2! n! 1 1 1 a1 = 1 + 1, a 2 = 1 + 1 + , a 3 = 1 + 1 + + etc.) 2 2 6 It is known that a n is convergent and lim a n = e .
n
Now for all n such that n 2, a n 1 + 1 + Therefore, e 2.5 Exercise: Show that e 2.65. We give here some standard limits: (1)
1 = 2.5. 2
Suppose r is a real number and r is a constant. (i) If r < 1 (i.e., -1 < r < 1), then lim r n = 0
n
lim n = 1
n
1 n
We end this section by giving a result that could be useful. Result: Suppose a n is a sequence. Then, lim a n = 0 if and only if lim a n = 0 .
n
Example: Let a n =
(1) n 1 for all n N. Then, a n = and hence lim a n = 0 . n n n n (1) Therefore, lim a n = 0 . i.e., lim = 0. n n n
We see that a1 a 2 a 3 a 4 ...... i.e., for all n, a n a n +1 . The terms of this sequence are non-increasing.
1 1 1 , , , ....... 2 3 4
We see that b1 > b2 > b3 > i.e., for all n, bn > bn +1 . The terms of this sequence are decreasing. Definition 1: Suppose a n is a sequence. Then: (1) (a) If for all n, a n a n +1 ,(i.e., a1 a 2 a 3 a 4 ...... ) we say that a n is monotonic decreasing (m.d). (b) If for all n, an > an+1 (i.e., a1 > a2 > a3 > ), we say that a n is strictly monotonic decreasing. (2) (a) If for all n, a n +1 a n (i.e., a1 a 2 a 3 a 4 ...... ) we say that a n is monotonic increasing (m.i). (b) If for all n, an+1 > an (i.e., a1 < a2 < a3 < ), we say that a n is strictly monotonic increasing. Note: For real numbers x, y, x y means that x > y or x = y. So, we see that if a n is strictly m.i it is also m.i and if a n is strictly m.d it is also m.d. We say that a sequence
an
a1 a 2 a 3 a 4 ...... or a1 a 2 a 3 a 4 ...... )
(2) Consider the sequence (1) n , i.e., -1, 1, -1, 1, This sequence is neither m.i. nor m.d. (3) Consider the sequence given recursively by a n +1 =
an for all n N. This 1 + an sequence is given, once the value of a1 is given. Then, by the recurrence equation, we get all the terms. Let a1 have a value such that a1 > 0. Then we see that for all
Question: What happens when a1 = 0? Definition 2: Suppose a n is a sequence and n0 N and n0 is a constant. Then: (1) (a) If for all n such that n n0 , a n a n +1 (i.e., a n0 a n0 +1 a n0 + 2 .......) , we say that a n is eventually monotonic decreasing. (b) If for all n such that n n0 , a n > a n +1 (i.e., a n0 > a n0 +1 > a n0 + 2 > .......) we say that a n is eventually strictly monotonic decreasing. (2) (a) If for all n such that n n0 , a n a n +1 (i.e., a n0 a n0 +1 a n0 + 2 .......) , we say that a n is eventually monotonic increasing. (b) If for all n such that n n0 , a n < a n +1 (i.e., a n0 < a n0 +1 < a n0 + 2 < .......) we say that a n is eventually strictly monotonic increasing. Note: The note given in definition 1 applies here too. Example: Consider the sequence
1 1 1 1 , i.e., -1, 1, , , , ......... 2n 1 5 7 9
1 . We see that for all n such that n 2 (i.e., we can take n0 = 2 2n 1 1 here), a n > a n +1 . So, is eventually strictly monotonic decreasing. It 2n 1 is also, eventually monotonic decreasing.
Let a n =
2.1.6. Bounded Sequences Definition: Suppose a n is a sequence. Then: (1) (2) If for all n, an k where k is a real number constant, we say that a n bounded above. If for all n, an k where k is a real number constant, we say that a n bounded below. is is
10
(3)
If a n
is
bounded. We see that a n is bounded means for all n, k1 an k2 where k1, k2 are real number is bounded if and only if an k, where k is a real number
Examples: (1) Consider the sequence (1) n , i.e., -1, 1, -1, 1, . For all n, -1 (-1)n 1. Hence, (1) n is bounded. Also, for all n, (1) n 1. (2) Consider the sequence n , i.e., 1, 2, 3, 4, . For all n, n 0 (actually n 1) Therefore, n is bounded below. However, n is not bounded above. Hence n is not bounded. (3) Consider the sequence n , i.e., -1, -2, -3, For all n, -n 0. Hence, n is bounded above. However n is not bounded below and hence it is not bounded. (4) Let us reconsider example (3) of 2.1.5, i.e., the sequence a n recurrence equation a n +1 = given by the
an and the value of a1. Consider when a1 0. 1 + an an 0. Then, for all n, an 0 and 1 1 + an Let K = max {a1, 1} (i.e., K is the greatest value in the set {a1, 1}). Then 0 an K for all n. Therefore, a n is bounded.
2.1.7. Relationship between monotonicity and boundedness Theorem: Suppose a n is monotonic (or eventually monotonic). (1) If
is
11
(2) If
n
an
is m.i (or eventually monotonic) and not bounded above, then, is m.d (or eventually monotonic) and bounded below, then, a n is
lim a n = .
(3) If a n
convergent. (4) If a n is m.d (or eventually monotonic) and not bounded below, then,
lim a n = .
n
In (1), if for all n such that n no, an k where no N and k R, then lim a n k (see theorem (4) in 2.1.4.)
n
In (3), if for all n such that n no, an k where no N and k R, then lim a n k (see theorem (4) in 2.1.4.)
n
the
sequence
an
where
for
all
n,
an =
2 2( n + 1) 2 . = 2 n +1 n +1 2 As n increases, decreases and hence an increases. Therefore, a n n +1 2 Also, for all n, a n = 2 2 . Therefore, a n is bounded above. n +1 Therefore, a n is convergent.
2n . n +1
is m.i.
1 ) = 20 = 2. n +1
is m.i
and not bounded above. lim a n = lim n 2 = . (3) Consider the sequence a n where for all n, an =
1 . n is m.d and bounded below (for all n, an 0 ).
an
1 = 0. n
(4) Consider the sequence a n where for all n, an = -n. Then, a n is m.d but not bounded below.
12
lim an = lim( n) = .
n n
(5) Let us consider the sequence in example 3 of 2.1.5 and let us take a1 > 0. We saw that then, a n is m.d. In example 4 of 2.1.6, we saw that a n is bounded above. Therefore a n is convergent. So, lim an = l for some l R . Let us find the value of l.
n
13
2.2
2.2.1
Infinite Series
Definition: Suppose a n is a sequence. We form the sequence Sn where for
n N , S n = a1 + a2 + .... + an = ai .
i =1
We say that Sn is the sequence of partial sums and Sn is the sum of the first n terms of the sequence a n . S1 = a1, S2 = a1 + a2, S3 = a1 + a2 + a3 etc. We call the sequence Sn a series and it is denoted by S1, S2, S3 etc are called the terms of the series. Example: Consider the sequence a n where an = Consider the series
1 . 2n 1
a
n =1
a
n =1
, i.e.,
2
n =1
1
n 1
2.2.2
a
n =1
Sn
a
n =1
a
n =1
is convergent).
If
Sn
a
n =1
a
n =1
is divergent).
a
n =1
= l.
Examples: 14
2
n =1
1
n 1
(see 2.2.1)
1 1 1 1 = 1 + + .... + n 1 = 2 n 1 . i 1 2 2 2 i =1 2 1 1 Now lim S n = 2 (since lim n 1 = lim( ) n 1 = 0 ). n 2 n 2 n 1 Therefore, n 1 is convergent and it converges to 2. n =1 2
a
n =1
1 . n(n + 1)
n 1 1 1 1 for any n N . = . Let S n = n(n + 1) n n + 1 r =1 r ( r + 1) n n 1 1 1 n 1 Then, S n = ( ) = r +1 r =1 r r =1 r r =1 r + 1 1 1 1 1 1 1 = (1 + + + ... + ) ( + + .... + ) 2 3 2 3 n n +1 1 = 1 n +1 (Note: We can also get this result in the following way: n n 1 n 1 1 n +1 1 1 ) = = 1 r r =1 r + 1 r =1 r r =2 r n +1 r =1 1 Therefore, lim S n = 1 (since, lim =0) n n + 1 n 1 Therefore, is convergent and it converges to 1. n =1 n ( n + 1)
a
n =1
Let S n = a1 + a 2 + ..... + a n , for any n. Then, Sn = 1 1 + 1 - . , to n terms, So, Sn = 1 1 + 1 1 + . + 1 1 = 0 when n is even, and Sn = 1 1 + 1 1 + . + 1 1 + 1 = 1 when n is odd. Therefore, S n is divergent. Therefore,
an is divergent, i.e.,
n =1
(1)
n =1
n +1
is divergent.
a
n =1
15
The sequence of partial sums S n is given by S n = 12 + 2 2 + ... + n 2 . For any n, S n = 12 + 2 2 + ... + n 2 n 2 . i.e., for any n, Sn n 2 and lim n 2 = .
n
Therefore, lim S n = .
n
n
n =1
is divergent.
The sequence of partial sums S n is given by, Sn = - 12 - 22 - - n2. Therefore for any n, S n n 2 and lim n 2 = .
n
Therefore lim S n = .
n
n
n =1
is divergent.
Note: A series
a
n =1
(ii) lim S n =
n
(In this case we say that the sequence S n is oscillatory) Possibility (i) is found in Example (4). Possibility (ii) is found in Example (5). Possibility (iii) is found in Example (3). 2.2.3 Fundamental Facts about Infinite Series
a
n =1
and
b
n =1
. Let c, 1 , 2 R be constants.
an and
n =1
b
n =1
are convergent.
16
(1)
ca n is convergent and
n =1
ca n = c an .
n =1
n =1
(2)
(a
n =1
+ bn ) is convergent and
n
(a
n =1
+ bn ) =
n
a +b
n =1 n
n =1
.
n
(3)
(1a n + 2 bn ) = 1
n =1
a n + 2 bn
n =1 n =1
(Note: This follows from (1) and (2). Also, when 1 = 1 and 2 = 1 , we get,
(a n bn ) =
n =1
an n =1
b
n =1
).
Theorem 2: Suppose a n is a sequence of non-negative terms and S n is a sequence of partial sums. Then,
a
n =1
some constant k R). (Note: See the Theorem in 2.1.7) Examples: (1) Suppose t n is a sequence such that for all n, t n {0,1,2,3,4,...9} and a n
is the
tn . 10 n
a
n =1
(Note: This is actually the infinite decimal, 0.t1t 2 t 3 ..... ) Let S n be the sequence of partial sums. Then for any n, S n =
t t1 t 9 9 9 + 2 2 + ... + n n + 2 + ..... + n . 10 10 10 10 10 10
17
1 1 n 9 9 9 9 1 1 9 1 + 2 + ..... + n = (1 + + ...... + n1 ) = ( 10 ) = 1 n 1 . 10 10 10 10 10 10 10 1 1 10 10 Therefore, for all n, Sn 1; i.e., S n is bounded above. Also, a n is a sequence of non-negative terms and hence S n is m.i. Therefore,
a
n =1
is convergent, i.e.,
10
n =1
tn
n
is convergent.
10
n =1
tn
n
where a n =
is a sequence of non-negative
=1+
n 2
Therefore, S 2 n +1 1 1 +
n
n =1
is divergent.
a
n =1
is divergent.
a
n =1
is convergent, then
18
a
n =1
is
This can be so in any one of the following four ways. (i) lim a n = l for some l such that l 0 .
n
(ii) lim a n =
n
(iii) lim a n =
n
ar
n =1
n 1
(i.e., a + ar + ar 2 + .....)
n , r
n 1
)
n n 1
= .
ar
n =1
n 1
is divergent.
n 1
=0.
n 1
ar
n =1
is divergent.
when n is odd a ar n 1 = a (1) n 1 = when n is even a Therefore, neither ar n -1 is convergent, nor lim ar n 1 = , nor lim ar n 1 = .
n
n 1
=0.
19
ar
n =1
n 1
is divergent.
Therefore, not lim ar n 1 = 0 (see 2.1.4, Theorem 4) Therefore, not lim ar n 1 = 0 ( see the last result in 2.1.4).
n
ar
n =1
n 1
is divergent.
This brings us to the final possibility. When -1 < r < 1: lim r n = 0 (see 2.1.4)
n
Therefore, lim ar n 1 = 0 .
n
ar
n =1
n 1
is convergent.
Let S n be the sequence of partial sums. a (1 r n ) a a = rn 1 r 1 r 1 r a Therefore lim S n = . n 1 r Sn = Therefore, when-1< r < 1,
n =1
a 0 as n , since rn 0 as n . 1 r
ar n1 is convergent and
ar
n =1
n 1
a . 1 r
n
n =1
1 1 is divergent. (i.e., is divergent). n n =1 n =1 n Let us consider the situation for the other values of .
=1,
When < 0: = -k where k = - > 0. 1 = n k an lim n k = . ( lim n k = when k > 0 is a standard limit). n n n 1 1 Therefore, lim = and hence by the theorem, is divergent. n n n =1 n
20
Therefore,
n
n =1
When = 0: 1 1 = 1 and hence lim = 1 . n n n 1 Therefore, not lim = 0 . n n 1 Therefore, is divergent when = 0. n =1 n When 0 < < 1: We will postpone the consideration of this case. We will consider it after the next theorem (i.e., Theorem 4). When > 1:
1 1 = 0 , this does not ensure that is convergent. However, it n n n =1 n is in fact convergent and we shall now show this: Let S n be the sequence of partial sums.
Although lim
Let n N .
1 1 1 1 1 1 1 1 1 + ) + ( + + + ) + .... + ( n + n + ... + n +1 ) (2 1) (2 ) ( 2 + 1) 4 5 6 7 2 3 1 1 1 1 + ( 2) + ( 4) + ..... + ( n 2 n ) 2 4 (2 ) 1 1 1 = 1 + 1 + 1 2 + .... + 1 n (2 ) 2 (2 ) 1 = 1 + r + r 2 + .... + r n where r = 1 and 0 < r < 1. (as 1 > 0) 2 1 r n +1 = 1 r 1 r n +1 1 = 1 r 1 r 1 r 1 1 Therefore, S 2 n +1 1 where r = 1 . 1 r 2 S 2 n +1 1 = 1 + (
Let m N . Then, there is n N such that 2 n +1 1 > m (since, lim(2n +1 1) = ). Since the terms of the series are non-negative and
n
2 n +1 1 > m , S m S 2 n +1 1
1 . 1 r
21
n
n =1
In the next two theorems we will consider only series of positive terms (i.e., series
a
n =1
an ,
n =1
b
n =1
bn is convergent, then,
n =1
a
n =1
is also convergent.
a
n =1
is divergent, then,
b
n =1
is
We said in Example 2 of Theorem 3 that we will be considering here the series when 0 < < 1. Let 0 < < 1.
n
n =1
1 1 (since, n n when 0 < < 1) n n 1 1 1 But is divergent. Therefore, is divergent; i.e., is divergent when n =1 n n =1 n n =1 n 0 < < 1. So, we have from this and what we had in Example 2 of Theorem 3, that,
n
n =1
22
an ) = l for some l R . n b n =1 n =1 n a (Note: By Theorem 4 of 2.1.4, we have that l 0 since for all n N , n > 0 ). bn Then:
Let
an ,
If l > 0,
b
n =1
is
convergent,
a
n =1
is also convergent.
x
n =1
a
n =1
1 212
1 . Now consider the series bn n n =1 1 = 100 and for the other values of n, 2
b
n =1 n =1
a
n =1
, i.e.,
as that of the original series. Let S n be the sequence of partial sums of the original series and let Tn
sequence of partial sums of the new series. Then, Tn = S n + k for all n such that n n0, where k R and n0 N are constants. On the other hand, if we delete a finite number of terms (for example consider a series
a
n =1
b
n =1
a3 and a14), with the notation used above we have, Tn = S n + n0 + k for all n such that n n1, where k R and n0 , n1 N are constants. If both these changes are made, we get Tn in the above form. In all these, the convergence or divergence of the new series is the same as that of the original series. * So, although we insisted in Theorems 4 and 5, that for all n, an > 0 and bn > 0, the conclusions of these two theorems still remains true if, for all n such that n n0, an > 0 and bn > 0 where n0 N is a constant. Do the exercises on Page 276 of Ref 1.
23
Alternating series: A series whose terms are alternately positive and negative is said to be an alternating series. An alternating series is of the form an > 0. Theorem 6 (Alternating Series Theorem): Let
(1)
n =1
n +1
(1)
n =1 n +1
n +1
an
Then,
(1)
n =1
a n is convergent.
(1)
n =1 n =1
n +1
a n where
n +1
an
(1)
1 is n
a
n =1
a
n =1
is
a
n =1
is convergent.
Definition: A series is said to be conditionally convergent if it is convergent but it is not absolutely convergent. Examples:
1 1 1 1 . ( 1) n +1 = and so ( 1) n +1 n n n n n =1 n =1 1 1 is which is divergent. Therefore ( 1) n +1 is conditionally convergent. n n =1 n n =1 (See the above example) 1 (2) The series ( 1) n +1 2 is absolutely convergent. n n =1
(1)
n +1
24
a
n =1
all n n0 , a n 0 where n0 N is a constant). Then: (1) If lim (2) If lim (3) If lim
a
n =1
a
n =1
a n +1 =, n a n
a
n =1
is divergent.
a
n =1
a
n =1
(2) If lim a n
n
a
n =1
(3) If lim a n
n
a
n =1
is divergent.
a
n=0
where an =
a
n =1
appearing just after Theorem 5). 1 a a 1 n! (n + 1)! Then, n +1 = = = 0 as n . (i.e., lim n +1 = 0) . Since n a 1 (n + 1)! n + 1 an n n! 1 1 0 < 1, by the Ratio Test, is convergent and hence, is convergent. n =1 n! n = 0 n! (Note: (i) It is absolutely convergent, but since the terms are positive, it is more sensible to say that it is convergent. 1 1 1 1 1 1 (ii) = 1 + 1 + + + .... , while = 1 + + + .... ) 2! 3! 2! 3! n = 0 n! n =1 n!
25
* Note:
logarithm of x and it is written as ln x. In this context, we also mention another standard limit (see 2.1.4), namely 1 lim (1 + ) n = e . n n We also considered this series in 2.1.4 as an example for Theorem 4 and we showed there that e 2.5. (2) Consider the series
a
n =1
n , where a n =
nn . n!
(n + 1) n +1 a (n + 1) n 1 (n + 1)! Then, n +1 = = = (1 + ) n . n n an n n n n! 1 Now, we know that, lim (1 + ) n = e 2.5 . n n a Therefore, lim n +1 = e > 1 . n a n
nn Therefore, is divergent. n =1 n! Note: This can be more easily proved using the Divergence Test (Theorem 3) as nn n 1 for all n. n = n 4 ..... n 1 2 ..... n = n ! for all n and hence n 244 14 3 n! n factors
n2 a n for any n +1
a
n =1
is divergent.
n 2 n , i.e., the series n =1
a
n =1
where a n =
n . 2n
26
1 n
1 n
is convergent.
Exercise: Use the Ratio Test to show that this series is convergent. (5) Consider the series
nn +1 2n 1 . n =1
nn +1 nn +1 nn n = for all n N. = 2n 1 2n 1 2n 2
n n n n n +1 Therefore, n = as n . 2 2 2 1
n 1 n 1
nn +1 2 n 1 is divergent. n =1
Use some of the steps in the above proof and with the use of the Comparison Test (Theorem 4), get the above result by using the Ratio Test.
e 2 n
1 n 1 n
1 n n
en Therefore, lim n 2n
Therefore,
en 2n is divergent. n =1 en n + 1 is divergent. n =1
27
In connection with determining whether a series is convergent or whether it is divergent, we may need a few more standard limits. We give these standard limits without going into much detail: (1) If lim x n = c and lim f ( x ) = l , then, lim f ( x n ) = l .
n x c n
(2)
(i) (ii)
(3)
Suppose r , s R and r, s > 0 and r, s are constants. Then, lim n r (ln n) s = ((ln n)-s is defined for all n such that n 2) and,
n
lim n r (ln n) s = 0 .
n
n for all n such that n 2 (see * (ln n) 5 n =1 remark, just after Theorem 5), is divergent.
where an =
*Exercise: Consider
an = 1 n
1 2
b
n =1
where bn =
b
n =1
is
ln(n + 1)
n =1
*Note: We take this occasion to belatedly state a result on sequences. Result on Sequences: Suppose a n is a sequence and k is an integer constant. Suppose
l R . Then:
28
Note: When k > 0, a n + k is the sequence bn where for any n, bn = a n + k . (i.e., b1 = ak+1, b2 = ak+2 etc). For instance, when k = 5, b1 = a6, b2 = a7, b3 = a8 etc). When k < 0, a n + k is a sequence bn where bn = a n + k when n k + 1 (i.e., n + k 1), and when 1 n k , bn could take whatever value we give them. For example, a n5 is a sequence bn where for n 6, bn = a n 5 (i.e., b6 = a1 , b7 = a2, b8 = a3 etc) and b1, b2, b3, b4 and b5 could take whatever value we give them. 1 and Tn ln(n + 1) is the sequence of partial sums of the sequence (i.e., T 1 = c1, T2 = c1 + c2, T3 = c1 + c2 + c3 etc) and S n is the sequence of partial sums of the sequence So, in the previous exercise, if c n is the sequence where cn =
bn ,
then, we have: Tn = 1 1 1 1 1 1 and S n = b1 + + + .... + + + .... + ln 2 ln 3 ln(n + 1) ln 2 ln 3 ln n and for any n such that n 2, Tn = Sn+1 b1, i.e., Tn = Sn+1 + constant.
From this we get, Tn is convergent if and only if S n is convergent. Since we showed that S n is divergent, we get that Tn is divergent, i.e., divergent.
ln(n + 1)
n =1
is
29
a
n =0
( x c) n
where a n is a sequence with first term a0 and x, c R and c is a constant while x is a variable. We say that, this is a power series about c. An important special case is when c = 0 where we have the power series about 0 which is
a
n=0
xn .
From a given context it is understood the number c about which it is a power series and we refer to any one of these series as just a power series. Example: Consider the series 1 + x + x 2 + x 3 + .... . This is the power series
a
n=0
xn
where an = 1 for all n N {0} . It is also a geometric series and it is convergent only 1 when x < 1 and 1 + x + x 2 + x 3 + .... = when x < 1, i.e., -1 < x < 1. 1 x Theorem 1: Consider a power series following possibilities occur: (1) It converges for all values of x in R. (2) It converges only when x = c. (3) There is a real number R1 , such that the power series converges for all x R such that c R1 < x < c + R1 (i.e., x c < R1 ) and diverges when x > c + R1 and when
a
n =0
30
xn n! . By application of the Ratio Test we get that the n =0 radius of convergence is (i.e., it converges for all x R )
(n + 1)
n=0
that the radius of convergence is 0 and the series converges only when x = 1.
xn x x2 x3 (3) Consider the power series n (i.e., 1 + + + + ..... ). By the Ratio Test 2 4 8 n =0 2 we get that the radius of convergence is 2. It does not converge when x = 2 and when x = -2. Therefore, the interval of convergence is (-2, 2), i.e., the set of all x R such that -2 < x < 2.
(4) Consider the power series 1 + x + all n 1 and a0 = 1. x2 x3 + + ..... , i.e., 2 3
a
n=0
x n where a n =
1 for n
In order to find the radius of convergence and the interval of convergence, it is xn sufficient to consider the series . Then, by the Ratio Test we get that the n =1 n 1 radius of convergence is 1 and when x = 1 the series diverges (since is n =1 n divergent) and when x = -1 the series converges by the Alternating Series Theorem (Theorem 6). (Note: Although in an alternating series the first term is positive, here the first term is negative. But this does not matter since, if we have a series
(1) n a n with
n =1
(1)
n =1
n +1
the sequence of partial sums of the series partial sums of the series diverge together. )
(1)
n =1
(1)
n =1
n +1
Therefore, the interval of convergence is [-1, 1), i.e., the set of all x R such that 1 x < 1.
31
a
n =0
that the radius of convergence is R1 for some R1 such that R1 > 0. Let I denote the interval of convergence. Let f be the function given by, f (x) = Then, for any x R such that, c R1 < x < c + R1,
f ( x) = na n ( x c) n 1 = a1 + 2a 2 ( x c) + 3a3 ( x c) 2 + ....
n =1
a
n =0
( x c) n , x I.
f is differentiable at x when x c < R1 (i.e., c R1 < x < c + R1) and f (x ) is equal to the series
na n ( x c) n1 (or
n =1
(n + 1)a
n=0
n +1
a
n =0
is differentiable at x for all x R, and the series also has as its radius of convergence.
na
n =1
( x c) n 1 which is equal to f (x )
(i.e., f k ( x ) = k!a k + ( 2 3 ... ( k + 1)) a k +1 ( x c) + (3 4 ... ( k + 2)) a k + 2 ( x c ) 2 + .... , dk dk dk (a k ( x c) k ) + k (a k +1 ( x c) k +1 ) + k (a k + 2 ( x c) k + 2 ) + .... ) dx dx dx k where x takes the values as it was for f (x ) . i.e., f k ( x) = *(3) All this applies when the radius of convergence is and here x takes all values in R, for f (x ) and for f k (x) where k N and all the series mentioned have radius of convergence . f k (c ) . *Exercise/Result: Show that for any k N , a k = k!
32
a
n =0
a n ( x c) n +1 + C where C is the integration constant and the radius of (a) n +1 n =0 convergence of this power series is also R1. f ( x)dx =
a1 a ( x c) 2 + 2 ( x c) 3 + ..... + C 2 3 n +1 a ( x c) +C) = n n +1 n =0
= a 0 ( x c) + (b) If I is the interval of convergence and a, b I and a < b and a, b are interior points of I, then,
a ( x c) n +1 a (b c) n +1 a n (a c) n +1 f ( x)dx = n = n n + 1 a n =0 n +1 n +1 n =0 a n (b c) n +1 a n (a c) n +1 = n +1 n +1 n =0 n =0
(Note:
b b b a n ( x c) n +1 = a n ( x c) n dx and hence, ( a n ( x c) n )dx = ( a n ( x c) n dx) ). a n +1 a a n =0 n =0 a b
*Note: When the radius of convergence is , f (x) = hold but in (a) the radius of convergence of
a
n =0
Examples: (1) f ( x) =
33
* This function is called the exponential function and f (x) is denoted by exp(x). Also, it is denoted by ex.
d x e = e x for all x R dx
So we have
(2)
(1)
n=0
1 x2 x3 x4 dx = x + + ...... + C where -1 < x < 1 1+ x 2 3 4 (Note: Actually this is true for 1 < x 1 ) n x + C , where C is the integration constant. = (1) n 1 n n =1
(Since,
1 + x dx = ln(1 + x) + constant )
3
When x = 0, we get, ln 1 = K, i.e., K = 0. Therefore, ln(1 + x) = (1) n 1 xn x2 x = x + .... where -1 < x < 1 n 2 3 n =1 (Note: Actually this is true for 1 < x 1 )
a
n =0
34
Suppose, (i.e.,
a
n =0 n
( x c) n is convergent when x = c + R1
a
n=0
( lim f ( x ) means the limit of f(x) as x tends to b but x taking values such that x < b)
x b
Example: x2 x3 x4 We saw that, ln(1 + x) = x + + ... where -1 < x < 1. 2 3 4 By the theorem, 1 1 1 lim ln(1 + x) = 1 + + .... x 1 2 3 4 1 1 1 i.e., ln 2 = 1 + + .... . 2 3 4
a
n=0
Definition: Now, let us consider functions, not given in terms of power series (example f(x) = sin x, x R )
35
f n (c ) ( x c) n the Taylor Series for f about c and we call the Taylor n! n =0 Series for f about 0 the Maclaurin Series for f. (Note: We must have that f is infinitely differentiable at x = c)
We call the series
We give without much details (Note: These details are required in order to apply the theorem) the following theorem: Theorem: Suppose f is a function defined on an interval I (Note: I could be R) and c is an interior point of I. f n (c ) ( x c) n for all x such that c r1 < x < c + r1 for some r1 > 0 (or Then: f ( x) = n! n =0 for all x R ), i.e., f(x) is equal to its Taylor Series for all x such that c r1 < x < c + r1 for some r1 > 0 (or for all x R ). f n (c ) n x (i.e., f(x) is equal to its Maclaurin Series) for all x such When c = 0, f ( x) = n! n =0 that r1 < x < r1 for some r1 > 0 (or for all x R ). Note: Though we have not given the details regarding the function f that are necessary for the theorem, we give the following facts. (1) Obviously, f (x) can be differentiated infinitely at x = c (i.e., f n (c) exists for all n N) f n ( y) (2) For any x, y (c r1 , c + r1 ), lim ( x c) n = 0 (or for any x, y R, n n! f n ( y) lim ( x c) n = 0 ). n n! This is a sufficient but not necessary requirement. Examples: (1) f (x) = sin x, x R . Let x R . Then f ( x ) = cos x, f ( x) = sin x, f ( x ) = cos x , and f ( x) = sin x . Therefore for any n N, f 4 n 3 ( x) = cos x, f 4 n 2 ( x) = sin x, f
4 n 1
( x) = cos x and f
4n
4n
( x) = sin x , i.e., f
4 n 3
(0) = 1, f
n
4n2
(0) = 0, f
4 n 1
(0) = 1
f ( y) n x = 0 (since f n ( y ) 1 and n!
x x f n ( y) x n and lim = 0 ). n n! n! n!
n =1
n!
n!
= 0 ).
36
(2) f ( x) = (1 + x) r where -1 < x < 1. Here, r R but r N and r 0. Let -1 < x < 1. Then, f ( x) = r (1 + x) r 1 , f ( x) = r (r 1)(1 + x) r 2 , etc., i.e., f n ( x) = r (r 1)......(r n + 1)(1 + x) r n . Then the Maclaurin Series
n =0
f n ( 0) n x is n!
r (r 1)....( r n + 1) n x . n! n=0
This series is called a binomial series. For further examples/problems, see Ref 5, pages 433 - 439, examples 2, 3, 7 and 8 and problems 1 3 (for problem 3 see example 8 on page 423), 4, 5 and 9-17. To solve some of the problems we state a useful result. Result: Suppose is an interior point of the interval of convergence of is an interior point of the interval of convergence of
( a n n ) ( bn n ) is equal to the convergent series
n N {0},
n =0 n =0
a
n=0 n n
x n and
b x
n =0 n
Then,
c
n=0
(i.e., ( a 0 + a1 + a 2 2 + a 3 3 + ...)(b0 + b1 + +b2 2 + b3 3 + ...) = a 0 b0 + (a 0 (b1 )) + ( a1 ) b0 ) + ( a 0 (b2 2 ) + ( a1 )(b1 ) + ( a 2 2 )b0 ) + ..... ) When = , we get, c n = n (a 0 bn + a1bn 1 + a 2 bn 2 + ... + a n b0 ) and
( a 0 + a1 + a 2 2 + a 3 3 + ...) (b0 + b1 + +b2 2 + b3 3 + ...)
= a 0 b0 + (a 0 b1 + a1b0 ) + 2 ( a 0 b2 + a1b1 + a 2 b0 ) + 3 ( a 0 b3 + a1b2 + a 2 b1 + a 3 b0 ) + *Note: This result could sometimes be used to justify that a function f is equal to its Maclaurin series (we need also the Ex/Result given just before Theorem 3) ------------------------------------------------- END ---------------------------------------------37