Stochastic integral representations and classification of sum- and max-infinitely divisible processes (Q5963496)

The paper deals with sum- and max-infinitely divisible processes. It consists of five sections and 46 references. Section 1 gives basic definitions and a history of the development of stochastic integral representations of infinitely divisible processes. In Section 2, minimal spectral representations for max-i.d.\ processes are introduced. Then, existence and uniqueness results for such representations under the condition (called Condition S) of separability of probability are formulated. Moreover, parallel results for i.d.\ processes are given. Next, necessary and sufficient conditions for the existence of measurable versions of max-i.d.\ and i.d.\ processes in terms of their spectral representations are established. Section 3 deals with connecting stationary i.d.\ and max-i.d.\ processes with measure preserving flows. The authors extend the theory of stable and max-stable processes by obtaining flow representations and ergodic decompositions for stationary i.d.\ and max-i.d.\ processes. Next, the authors prove that the minimal flow representation associated with a stationary stochastically continuous max-i.d.\ or i.d.\ process is unique up to a flow isomorphism. In Section 4, several examples and applications are supplied. Section 5 contains the proofs of several lemmas and theorems. The proofs are written clearly with all details. This long paper (36 pages!) finishes with 46 references which are representative for the studied problems. The paper seems to be interesting for people working on infinitely divisible processes.