On the density of the supremum of a stable process

TL;DR

This paper investigates the density of the supremum of stable Lévy processes, revealing that series representations are not always absolutely convergent for irrational stability parameters and providing explicit formulas for rational cases.

Contribution

It demonstrates the non-universality of series convergence for irrational parameters and derives explicit Mellin transform formulas for rational stability indices.

Findings

Series do not always converge absolutely for all irrational α.

Explicit Mellin transform formulas involve Gamma functions and dilogarithms.

Numerical experiments illustrate theoretical results.

Abstract

We study the density of the supremum of a strictly stable L\'evy process. As was proved recently in F. Hubalek and A. Kuznetsov "A convergent series representation for the density of the supremum of a stable process" (Elect. Comm. in Probab., 16, 84-95, 2011), for almost all irrational values of the stability parameter $\alpha$ this density can be represented by an absolutely convergent series. We show that this result is not valid for all irrational values of $\alpha$: we construct a dense uncountable set of irrational numbers $\alpha$ for which the series does not converge absolutely. Our second goal is to investigate in more detail the important case when $\alpha$ is rational. We derive an explicit formula for the Mellin transform of the supremum, which is given in terms of Gamma function and dilogarithm. In order to illustrate the usefulness of these results we perform several numerical experiments and discuss their implications. Finally, we state some interesting connections that this problem has to other areas of Mathematics and Mathematical Physics, such as q-series, Diophantine approximations and quantum dilogarithms, and we also suggest several open problems.