Self-similar extremal processes (Q1286610)

Let \(Y=Y(0)\) and \(U(s,t)\), \(0\leq s<t\), be independent \(d\)-dimensional random variables. The point process \(Y(t)=\max(Y(s),U(s,t))\), \(0\leq x<t\), is called a \(d\)-dimensional max-extremal process with independent max-increments. The vectors \(U(s,t)\) are assumed to have a lower bound \(C(t)\) with some smoothness requirements. All operations are meant component by component. The author gives a characterization of all possible limiting processes \(Z(t)\) for \(Y_n(t)\), where \(Y_n(t)\) is a time and space normalized version of \(Y(t)\), and the normalization is assumed to achieve asymptotic negligibility.