Selfdecomposable fields (Q521968)

The authors use definitions of infinitely divisible and selfdecomposable random fields (random processes with `multidimensional' time) which are similar to the definitions for random variables. More precisely, a random field $X$ is called infinitely divisible, if, for each $n\in\mathbb{N}$, there exist independent and identically distributed random fields $X^{1,n},\ldots, X^{n,n}$ such that $X$ has the same distribution as $X^{1,n}+\ldots+X^{n,n}$. A random field $X$ is called selfdecomposable, if for each $q>1$ there exists an infinitely divisible random field $X^q$ independent of $X^\prime$ a copy of $X$ such that $X$ has the same distribution as $q^{-1}X^\prime+X^q$. \par The authors develop necessary theoretical background for selfdecomposable random fields. In particular, they define `master' Lévy measures, present a Lévy-Itô representation and prove a criterion for selfdecomposability of random fields in terms of dilations. Volterra fields are given by integrals in which the integrands are certain deterministic functions, called the kernel functions, and the integrators are so called Poisson Lévy bases. It is shown that under some conditions on the kernel selfdecomposability of the Volterra field is equivalent to selfdecomposability of the Lévy basis. This equivalence is shown to be valid with respect to Urbanik's classes. Also, the authors investigate `existence and selfdecomposability of integrated Volterra fields'; `introduce infinitely divisible field-valued Lévy processes; give the Lévy-Itô representation associated with them; study stochastic integration with respect to such processes'; `provide examples in the form of Lévy semistationary processes with a Gamma kernel and Ornstein-Uhlenbeck processes'.