Regularizing mappings of Lévy measures (Q2490069)

One introduces a mapping \(\Upsilon_{0}\) which is one-to-one from the class of one-dimensionnal Lévy measures into itself. The authors investigate a mapping \(\Upsilon\) which is one-to-one from the class of one-dimensionnal infinite divisible probability measures into itself. Based on the study of \(\Upsilon_{0}\) one proves further properties of \(\Upsilon\), for instance it maps the class of selfdecomposable laws onto the so-called Thorin class (the smallest subclass of infinitely divisible probability measures on \(\mathbb R\) which contain all positive and negative Gamma distributions and is closed under convolution and weak limits). The authors also introduce a one-parameter family \((\Upsilon^{\alpha})_{\alpha\in[0,1]}\) of one-to-one mappings from the class of infinitely divisible probability measures into itself, which interpolates smoothly between \(\Upsilon\) (\(\alpha=0\)) and the identity mapping (\(\alpha=1\)). One proves that each mapping \(\Upsilon^{\alpha}\) is representable in terms of stochastic integrals with respect to a Lévy process. The Mittag-Leffler function and law play a key role in the study.