Random Locations, Ordered Random Sets and Stationarity

Abstract: Intrinsic location functional is a large class of random locations containing locations that one may encounter in many cases, e.g., the location of the path supremum/infimum over a given interval, the first/last hitting time, etc. It has been shown that this notion is very closely related to stationary stochastic processes, and can be used to characterize stationarity. In this paper the author firstly identifies a subclass of intrinsic location functional and proves that this subclass has a deep relationship to stationary increment processes. Then we describe intrinsic location functionals using random partially ordered point sets and piecewise linear functions. It is proved that each random location in this class corresponds to the location of the maximal element in a random set over an interval, according to certain partial order. Moreover, the locations changes in a very specific way when the interval of interest shifts along the real line. Based on these ideas, a generalization of intrinsic location functional called "local intrinsic location functional" is introduced and its relationship with intrinsic location functional is investigated.