Some limit theorems on the increments of \(\ell^p\)-valued multi-parameter Gaussian processes (Q2577667): Difference between revisions

Let \(Y(t)=(X_1(t),X_2(t),\dots)\), \(t\in[0,\infty)^N\), \(N\geq 1\), be an \(\ell^p\)-valued multiparameter centered Gaussian process with coordinate processes continuous and mutually independent. Suppose each \(X_k(t)\) has stationary increments with variance satisfying \(\sigma_k^2(\|{\mathbf h}\|)=E(X_k(t+{\mathbf h})-X_k(t))^2\). Some limit theorems for the norm of increments \(\|Y(t+s)-Y(t)\|_{\ell^p}\) are proved under various conditions on \(\sigma_k(h)\). These conditions are weaker than those assumed by \textit{M. Csörgő} and \textit{Q.-M. Shao} in the one-parameter case [Ann. Probab. 21, No. 4, 1958--1990 (1993; Zbl 0791.60028)].