Saturday, 20 May 2017

calculus - Limit of a convergent series



For a research project, after some manipulation I come up with a convergent series that I have to prove its limit. The statement is the following:



$ \lim_{n \rightarrow \infty } \displaystyle \sum_{k=1}^n \left[ \prod_{j=n-k+1}^\infty (1-q^j)-1 \right]q^k = 0, \; 0

The LHS can be rewritten by the q-shifted factorial: (q,q)n=ni=1(1qi) as follows:



$\lim_{n \rightarrow \infty} \displaystyle \sum_{k=1}^n \left[ \frac{(q,q)_\infty}{(q,q)_{n-k}} -1 \right] q^k = 0, \; 0


In fact the second statement is the original limit that I have to prove. I tested the limit numerically by varying n, the series seem to converge pretty fast. However, I couldn't figure out how to prove it analytically. If anyone can have an idea to share, I would be grateful.


Answer



In your sums, call the terms in brackets c(k,n). Let's think of c(k,n) as defined for all k,nN, with c(k,n)=0 for k>n. So we are dealing with k=1c(k,n)qk. Now |c(k,n)|2 for all k,n. And for fixed k,limnc(k,n)=0. Because k=1qk<, the dominated convergence theorem tells you



limnk=1c(k,n)qk=k=1[limnc(k,n)qk]=0.


No comments:

Post a Comment

real analysis - How to find limhrightarrow0fracsin(ha)h

How to find limh0sin(ha)h without lhopital rule? I know when I use lhopital I easy get $$ \lim_{h\rightarrow 0}...