PDA

View Full Version : Convergence in "mean square" (or L^2) sense


kingwinner
Nov9-09, 10:10 PM
1. The problem statement, all variables and given/known data
This is an example from a textbook (with solutions) in which I am feeling confused.
Let fn(x) = [n/(1+n2x2)] - (n-1)/[1+(n-1)2x2] in the interval 0<x<L. This series telescopes so that
N
∑ fn(x) = N/(1+N2x2)
n=1

L
∫ [∑ fn(x)]2 dx =
0
L
∫ N2/(1+N2x2)2 dx =
0
NL
∫ N/(1+y2)2 dy (let y=Nx)
0
This last line -> +∞ as N->∞
Since it does not converge to 0, the series does NOT converge in the mean-square (or L2) sense to f(x)=0.

2. Relevant equations/concepts
Convergence in mean square/L2 sense

3. The attempt at a solution
N/A

(i) Now I don't understand why we have to use the change of variable y=Nx. What is the point of doing this?


(ii) Also, WHY as N->∞,
NL
∫ N/(1+y2)2 dy -> +∞ ?
0

Can someone please explain?
Thank you!

kingwinner
Nov10-09, 03:00 PM
Why is it a good idea to bring N into the limits of integration? After the change of variable, N appears both in the upper limit of integration AND in the integrand...