2
$\begingroup$

I am studying a chemical system in which a statistic has the following form:

$$ E \ = \ \frac{I_0(2\sqrt{\lambda})^2}{I_1(2\sqrt{\lambda})^2} - \frac{1}{2\sqrt{\lambda}} \frac{I_0(2\sqrt{\lambda})}{I_1(2\sqrt{\lambda})} - 1.$$

Where $I_0(x)$ is the modified Bessel Function of the First Kind and $\lambda \geq 0$. I have shown that $0 \leq E$ and I am trying to find an asymptotic upper bound for $E$ (expression for upper bound that depends on $\lambda$) but don't know how to do it because I don't have the math background for it! Any ideas on how this can be approached? Any help is appreciated.

Edit: Looks like $E \leq \frac{1}{2\lambda}$ from simulations. Might be wrong! Still, don't know if that is provable.

$\endgroup$
0
$\begingroup$

First, let $2 \sqrt \lambda=x$ to make $$E \ = \ \frac{I_0(x)^2}{I_1(x)^2} - \frac{1}{x} \frac{I_0(x)}{I_1(x)} - 1$$ and use the expansions $$I_0(x)=1+\frac{x^2}{4}+\frac{x^4}{64}+\frac{x^6}{2304}+O\left(x^8\right)$$ $$I_1(x)=\frac{x}{2}+\frac{x^3}{16}+\frac{x^5}{384}+\frac{x^7}{18432}+O\left(x^8\right)$$ which give $$E=\frac{2}{x^2}-\frac{1}{4}+\frac{x^2}{32}-\frac{5 x^4}{1536}+\frac{7 x^6}{23040}+O\left(x^8\right)$$ Back to $\lambda$ $$E=\frac{1}{2 \lambda }-\frac{1}{4}+\frac{\lambda }{8}-\frac{5 \lambda ^2}{96}+O\left(\lambda ^3\right)$$

$\endgroup$
  • $\begingroup$ I am a bit confused by this answer because when I use Mathematica, I get $E \leq \frac{1}{2\lambda}$ for $\lambda > \frac{1}{2}$. But this expression for $E$ would grow fast with large $\lambda$! Also, I know $\lim_{\lambda \to \infty} E = 0$ but this won't yield that either. $\endgroup$ – Analytic Potato Jul 16 at 12:07
  • $\begingroup$ @AnalyticPotato. As far as I understood, you are looking for tha asymptotics when $\lambda \to 0$ (for large values of $\lambda$ the answer was given in your othe question. $\endgroup$ – Claude Leibovici Jul 17 at 3:13

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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