Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

I have a question concerning the series:

$$c_n:= \sum_{k=0}^{\lfloor \frac{n}{2} \rfloor} \left(-a\right)^k \frac{1}{k!} \frac{b^{n-2k}}{\left(n-2k\right)!} ~ ~ ~ ~; ~ ~ ~ ~ a,b>0 ~ ~ ~ ; ~ ~ n\in\mathbb{N }$$

I could show, that $c_n$ converges to $0$ for $n$ to infinity, but I need a bound for the speed of decay (e.g. $\vert c_n \vert \leq 1/2^n$). If we destinguish between odd and even $n$ and apply wolfram alpha we get:

$$ c_n = \frac{4^n a^n ~ U(-n,1/2,b^2/4a)}{(2n)!} ~~~~\text{ ( n even )}$$ $$ c_n = \frac{4^n a^n ~ b ~ U(-n,3/2,b^2/4a)}{(2n+1)!} ~~~~\text{ ( n odd)}$$

Here $U$ denotes the confluent hypergeometric function. I know the definition of the confluent hypergeometric function as a series and I know that it is one of the solutions of Kummer's differential equation. Unfortunately I know nothing else about this topic. I look it up in the handbook of Abramowitz and Stegun and got the impression that there is a very big theory behind hypergeometric functions.

Is there maybe any elementary way to determine a good bound for the decay rate? The whole field of mathematics around the confluent hypergeometric function seems to be advanced and I guess it would take very long to get some basic knowledge. To be honest: all my standard textbooks do not cover this topic. By the way: is there any readable introduction to the field ?

With best regards, Mat

share|improve this question

1 Answer 1

up vote 1 down vote accepted

This is only a draft/sketch, which must be expanded at some stages. I think for your problem you should use formula 13.5.16 Abramowitz and Stegun to estimate the U values: $$ U(-n,1/2,b^2/4a) \sim \Gamma(\frac{1}{4} + n + \frac{1}{4})\Big(c_{even}(a,b)+O(1)\Big)$$ $$ U(-n,3/2,b^2/4a) \sim \Gamma(\frac{3}{4} + n + \frac{1}{4})\Big(c_{odd}(a,b)+O(1)\Big)$$ where $c_{odd}, c_{even}$ do not depend on $n$, but the $O(1)$ includes a $\cos(f\sqrt{n}+g)$ term. Plugging these estimates in your expressions for the $c_n$, Maple gets the asymptotic expressions $$ c_n = \frac{4^n a^n ~ U(-n,1/2,b^2/4a)}{(2n)!} \sim \frac{\sqrt{2}}{4 n}\left(\frac{ea}{n^{3/2}}\right)^n c'_{even}(a,b)~~~~\text{ ( n even )}$$ $$ c_n = \frac{4^n a^n ~ b ~ U(-n,3/2,b^2/4a)}{(2n+1)!} \sim \frac{\sqrt{2}}{4 n}\left(\frac{ea}{n}\right)^n c'_{odd}(a,b)~~~~\text{ ( n odd)}$$ which shows the decay rate for the $c_n$. But note: If had not made severe errors, the rate is much slower than $2^{-n}$.

share|improve this answer
    
hello and thank you for your answer. I will start working out the details :) Just one little question: did you mean that the rate of decay is much faster ? For some $n$ we have $ea/n \leq 1/2$... –  Mat Sep 17 '13 at 15:12
1  
You are right, I only looked at the $1/n$ factors. –  gammatester Sep 18 '13 at 7:48

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.