Some theorems concerning extrema of Brownian motion with \(d\)-dimensional time (Q5939262)

Let \(X(t)\), \(t\in R^d\), be a Lévy Brownian motion with \(d\)-dimensional time and let NEWLINE\[NEWLINEX_{*}(A)=\inf\{X(t): t\in A\}, \qquad X^{*}(A)=\sup \{X(t): t\in A\}.NEWLINE\]NEWLINE The author uses the notation \(X(A)\) to denote either \(X_{*}(A)\) or \(X^{*}(A)\). The paper presents some sufficient conditions on the sets \(A\) and \(B\) under which the probability distribution of \(X(A)\) or joint probability distribution of \(X(A)\) and \(X(B)\) admits a strictly positive density. For instance, the joint distribution of \(X(A)\) and \(X(B)\) has a strictly positive \(C^{\infty}\)-density provided that \(A\) and \(B\) are nonempty bounded closed sets separated from each other by a certain \((d-1)\)-dimensional hyperplane passing throw 0. Also it is proved that for almost all sample functions there are no distinct extreme points \(s\) and \(y\) such that \(X(s)=X(y)\).