Operator-self-similar stable processes (Q1343600): Difference between revisions

An \(\mathbb{R}^ d\)-valued stochastic process \(X(t)\), \(t \geq 0\), is called operator-self-similar if it is continuous in law, and for every \(c>0\) there exist a linear operator \(B_ c\) and a vector \(v_ c \in \mathbb{R}^ d\) such that the processes \(X(ct)\) and \(B_ c X(t) + v_ c\) have the same finite-dimensional distributions. Imposing the additional restriction that \(B_ c = cI\) \((I = \) identity operator) we get as a particular case self-similar processes. The authors construct some examples of operator-self-similar and operator-stable processes. These are nontrivial extentions of those processes for \(d = 1\). The monograph ``Operator-limit distributions in probability theory'' by the reviewer and the second author (Wiley, 1993) gives full account of limit distributions for partial sums normalized by linear bounded operators. As particular case, operator-stable laws are studied.