View Single Post
Old 10-25-2004, 04:48 PM   #5 (permalink)
phukraut
Addict
 
An example of molloby's idea appears on page 77 of Goldberg's Methods of Real Analysis (second edition):

We have that

(1) \sum_{n=1}^\infty (-1)^{n+1}/n = L = \log 2. (where L is the sum of the series).

Now, divide L by 2, giving

L/2 = 1/2 - 1/4 + 1/6 - 1/8 + ..., and thus

(2) L/2 = 0 + 1/2 - 0 - 1/4 + 0 + 1/6 - 0 - 1/8 + ... .

Adding (2) and (1) gives

(3) 3/2 L = 1 + 1/3 - 1/2 + 1/5 + 1/7 - 1/4 + ... .

You will notice that (3) is just a rearrangement of (1), yet it sums to 3/2 L = 3/2 \log 2.

The book goes on to give the following theorem:

Let L be a conditionally convergent series of real numbers. Then for any real number x, there is a rearrangement of L which converges to x. However, if the series is absolutely convergent, then all rearrangements sum to the same value.
phukraut is offline  
 

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37