Finding the Sum of an Alternating Series

In summary, you got the sum of 2009-2008+2007-2006+...+3-2+1 by grouping them using the equivalent series. Your answer is correct.
  • #1
mplayer
152
0

Homework Statement



Evaluate the sum [tex]2009[/tex][tex]^{2}[/tex] - [tex]2008[/tex][tex]^{2}[/tex] + [tex]2007[/tex][tex]^{2}[/tex] - [tex]2006[/tex][tex]^{2}[/tex] + ... + [tex]3[/tex][tex]^{2}[/tex] - [tex]2[/tex][tex]^{2}[/tex] + [tex]1[/tex][tex]^{2}[/tex]

Homework Equations



I think that the equivalent series representation of this sum is:

[tex]\sum^{2009}_{n=1}n^{2}(-1)^{n+1}[/tex]

The Attempt at a Solution



I vaguely remember in one of my calculus classes way back when something about finding the convergence of a series, I just don't remember how exactly to do it. I'm sure that is probably the easiest method to obtaining the sum. I found the sum by a somewhat roundabout method, which just consisted of finding patterns within patterns eventually reducing the 2009-part summation to a 12-part summation as follows:

435 + 21510 + 53910 + 86310 + 118710 + 151110 + 183510 +215910 + 248310 + 280710 + 313110 + 345510

resulting in a sum of 2,019,045.

Could someone who knows how to find the solution to the series please check my answer? Thanks!
 
Last edited:
Physics news on Phys.org
  • #2
Your closed-form series representation is incorrect. You have
[tex]\sum^{2009}_{n=1}n(-1)^{n+1}[/tex]
Notice that every term in the original problem is squared.

How did you get this?
eventually reducing the 2009-part summation to a 12-part summation as follows:

435 + 21510 + 53910 + 86310 + 118710 + 151110 + 183510 +215910 + 248310 + 280710 + 313110 + 345510
Is this based on the first representation? It might be right, but as you haven't explained how you got it, I have no way of telling whether it's correct short of figuring this out myself, which I am reluctant to do.

I don't believe the answer you got, 2,019,045, is correct. The 2009 term alone is larger than that by a factor of about 2.
 
  • #3
Ah yes, thanks, forgot to put that squared term in there. I changed it in my original post, is the series representation correct now?
 
  • #4
The way that I reduced it took a few steps.

First, I summed together adjacent terms to get positive numbers. This reduced the 2009+1-part summation(including the term 0^2 to make the number divisible by 2) to a (2009+1)/2 = 1005-part summation.

0^2 + 1^2 = 1
-(2^2) + 3^2 = 5
-(4^2) + 5^2 = 9
-(6^2) + 7^2 = 13
-(8^2) + 9^2 = 17

Notice how each term is just increasing by 4. So I rewrote this sequence to see if I could find another pattern to again reduce the number of terms to sum.

1,5,9,13,17,21,25,29,33,37,41,45,49,53,57,61,65,69,73,77

Notice how the first 5 terms sum is 45. The next 5 terms sum is 145. The next 5 terms sum is 245. The next 5 terms sum is 345...etc. This reduces the previous 1005-part summation to a (1005/5)=201-part summation. I again rewrote this sequence to see if I could find yet another pattern.

I kept using this method to finally reduce it down to a 12-part summation that I listed I my original post. It seems to make sense to me, but I'm really not sure which is why I was looking for a more direct method so I could check my answer.
 
  • #5
mplayer said:

Homework Statement



Evaluate the sum [tex]2009[/tex][tex]^{2}[/tex] - [tex]2008[/tex][tex]^{2}[/tex] + [tex]2007[/tex][tex]^{2}[/tex] - [tex]2006[/tex][tex]^{2}[/tex] + ... + [tex]3[/tex][tex]^{2}[/tex] - [tex]2[/tex][tex]^{2}[/tex] + [tex]1[/tex][tex]^{2}[/tex]

Homework Equations



I think that the equivalent series representation of this sum is:

[tex]\sum^{2009}_{n=1}n^{2}(-1)^{n+1}[/tex]

The Attempt at a Solution



I vaguely remember in one of my calculus classes way back when something about finding the convergence of a series, I just don't remember how exactly to do it. I'm sure that is probably the easiest method to obtaining the sum. I found the sum by a somewhat roundabout method, which just consisted of finding patterns within patterns eventually reducing the 2009-part summation to a 12-part summation as follows:

435 + 21510 + 53910 + 86310 + 118710 + 151110 + 183510 +215910 + 248310 + 280710 + 313110 + 345510

resulting in a sum of 2,019,045.

Could someone who knows how to find the solution to the series please check my answer? Thanks!

This doesn't have anything to do with convergence. It is just a sum of finitely many terms. Try grouping it like this:

[tex] 1^2 + (-2^2 + 3^2) + (-4^2 + 5^2) ... (-2008^2+2009^2)

= 1+\sum_{k=2}^{2008}(-k^2+(k+1)^2) = 1+\sum_{k=2}^{2008}(2k+1)[/tex]

The last expression is just an arithmetic series which you can sum. I didn't get your answer but it's late here so I'll let you take it from there.
 
  • #6
OK, now I understand, and your approach is good one, especially the part where you noticed that each successive pair of terms increases by 4. I get the same value you did for this series, 2,019,045.

Here is another explanation that uses your idea.

Let S = (-02 + 12) - (22) + 32) -+ ... - (20082 + 20092)

There are 1005 pairs of terms, which is easy to see if you index them starting from 0 and ending with 1004.

So,
[tex]S~=~\sum_{n = 0}^{1004} (1 + 4n)[/tex]
This gives is 1 + 5 + 9 + ...
We can evaluate the summation above by breaking it up into two summations.
[tex]S~=~\sum_{n = 0}^{1004} 1 + \sum_{n = 0}^{1004}4n[/tex]
[tex]=~\sum_{n = 0}^{1004} 1 + 4\sum_{n = 1}^{1004}n[/tex]
(Note that I brought the constant factor outside the 2nd summation and changed the lower limit on the same summation, which doesn't change the final value.)
= 1005 + 4*1005*502
= 2,019,045
 
  • #7
Thank you both very much for your help! :biggrin:
 

FAQ: Finding the Sum of an Alternating Series

What is a series solution check?

A series solution check is a mathematical technique used to verify the accuracy of a series solution to a differential equation. It involves comparing the terms of the series with the actual solution to determine the degree of precision.

When is a series solution check necessary?

A series solution check is necessary when solving a differential equation using a series expansion method, such as the power series or Frobenius method. It is important to ensure that the solution obtained is accurate and does not contain any errors.

How is a series solution check performed?

A series solution check is performed by substituting the series solution into the original differential equation and comparing the terms on each side. The terms should be equal for the solution to be considered accurate.

What are the advantages of using a series solution check?

Using a series solution check allows for the detection of errors or inaccuracies in the solution. It also helps to determine the radius of convergence of the series, which indicates the range of values for which the solution is valid.

Are there any limitations to series solution checks?

Series solution checks are limited to differential equations that can be represented as a power series or a Frobenius series. They may also not be applicable for non-analytic functions or for solutions with singular points.

Similar threads

Replies
6
Views
2K
Replies
4
Views
1K
Replies
16
Views
1K
Replies
22
Views
4K
Replies
16
Views
2K
Replies
14
Views
3K
Back
Top