Superextremal processes, max-stability and dynamic continuous choice (Q1336590)

This interesting paper deals with the analysis of a dynamic choice process \(M= \{M_ t : t \in (0,\infty)\}\), based on an infinite- dimensional extremal process (called superextremal) \(Y = \{Y_ t : t \in (0, \infty)\}\), i.e. \(M_ t = \{c \in \mathbb{E} : Y_ t(c) = \bigvee_{s \in \mathbb{E}} Y_ t (s)\}\) where \(\mathbb{E}\) is a compact metric space. Since the max-stable processes are very convenient for modeling the continuous choice problem, the most of the problems treated here concern superextremal processes with max-stable components \(Y_ t\) and their sup-Lévy measures. It is shown that the choice process is Markov, if \(Y\) has max-stable components. The stationarity of \(M\) in case of time- homogeneous \(Y\) is also checked. In the last section it is proved that for every superextremal process with max-stable components there is such one generated by a simple Poisson point process, so that both processes are equally distributed.