How to Expand Finite Series in a Shorter Form?

Join the discussion
Registration is free. Start your own thread to ask a follow-up.
4 replies · 2K views
JulmaJuha
Messages
2
Reaction score
0
Hello :blushing:

How to do expand this: [itex](\sum_{j=1}^{n}(X(t_j)-X(t_{j-1}))^2 - t)^2[/itex] where [itex]X(t_j)-X(t_{j-1}) = \Delta X_j[/itex]

to this: [itex](\sum_{j=1}^{n}(\Delta X_j)^4 + 2*\sum_{i=1}^{n}\sum_{j<i}^{ }(\Delta X_i)^2(\Delta X_j)^2[/itex] [itex]-2*t*\sum_{j=1}^{n}(\Delta X_j)^2+t^2[/itex]I get near the North Pole... but it seems that I've forgotten some fundamental rules of finite series to do the last part of the manipulation :smile:.I assume that the blue part has to be expaned further. This is what I've done:

[itex]\sum_{j=1}^{n}(\Delta X_i)^2 = a,t=b[/itex]

[itex]E[(a-b)^2]=E[a^2-2ab+b^2]=E[\color{blue} {(\sum_{j=1}^{n}(\Delta X)^2)^2} \color{black} -2*t*\sum_{j=1}^{n}(\Delta X)^2+t^2][/itex]

Any help would be greatly appreciated.
 
Last edited:
Physics news on Phys.org
I may be misunderstanding something, but I think the second term (with the double sum in [itex]i[/itex] and [itex]j[/itex]) should be multiplied by [itex]2[/itex]. Either that or the sum in [itex]j[/itex] should be over [itex]j \neq i[/itex] rather than [itex]j<i[/itex].

Anyway, you are correct to say that the blue term needs to be expanded further. Just try writing out one explicit example, say for [itex]n=2[/itex]. Often, the compact summation notation obscures otherwise obvious patterns.

Also, you can use [itex]( \sum_j f_j )^2 = ( \sum_j f_j ) ( \sum_i f_i )[/itex].
 
krome you're correct, its supposed to be multiplied by 2
 
JulmaJuha said:
krome you're correct, its supposed to be multiplied by 2

No, it should not be multiplied by 2. If ##a_i = (\Delta X_i)^2##, you have to expand the sum
[tex]\left(\sum_i (a_i-t)\right)^2 = \left( \sum_i a_i - nt \right)^2 <br /> = \sum_i a_i^2 + 2\sum_{i<j} a_i a_j - 2nt \sum_i a_i + n^2 t^2.[/tex]
 
JulmaJuha said:
[itex](\sum_{j=1}^{n}(\Delta X_j)^4 + 2*\sum_{i=1}^{n}\sum_{j<i}^{ }(\Delta X_i)^2(\Delta X_j)^2[/itex] [itex]-2*t*\sum_{j=1}^{n}(\Delta X_j)^2+t^2[/itex]

Good lord! I must be going mad or selectively blind. I swear when I read this last night the second term did not have a factor of 2! :confused: