Fox-H densities and completely monotone generalized Wright functions

TL;DR

This paper explores Fox-H densities, their properties, and their connections to generalized Wright functions, providing explicit conditions, subclasses, and asymptotic analysis relevant for probabilistic applications.

Contribution

It introduces a subclass of Fox-H densities with finite moments, characterizes their Laplace transforms as generalized Wright functions, and analyzes their existence and sign conditions.

Findings

Defined Fox-H densities with all moments finite

Identified eight relevant subclasses with probabilistic interpretations

Derived asymptotic and analytic extension results for Fox-H functions

Abstract

Due to their flexibility, Fox-$H$ functions are widely studied and applied to many research topics, such as astrophysics, mechanical statistic, probability, etc. Well-known special cases of Fox-$H$ functions, such as Mittag-Leffler and Wright functions, find a wide application in the theory of stochastic processes, anomalous diffusions and non-Gaussian analysis. In this paper, we focus on certain explicit assumptions that allow us to use the Fox-$H$ functions as densities. We then provide a subfamily of the latter, called Fox-$H$ densities with all moments finite, and give their Laplace transforms as entire generalized Wright functions. The class of random variables with these densities is proven to possess a monoid structure. We present eight subclasses of special cases of such densities (together with their Laplace transforms) that are particularly relevant in applications, thanks to their probabilistic interpretation. To analyze the existence conditions of Fox-$H$ functions red as well as their sign, we derive asymptotic results and their analytic extension.