Limsup results and LIL for finite dimensional Gaussian random fields (Q2460670)

The author considers a \(d\)-dimensional vector \(X^d(t)=(X_1(t),\dots,X_d(t))\), \(t\in R^N_+\), of independent continuous, isotropic and centered \(N\)-parametric Gaussian random fields. The main results are limsup and liminf theorems related to the law of iterated logarithm for the increments \(\| X^d(t+s)-X^d(t)\| \) when \(\| s\| \) and \(\| t\| \) tend to infinity under certain restrictions. The proofs are based on some inequalities for Gaussian vectors due to \textit{S. M. Berman} [Ann. Math. Statist. 35, 502--616 (1964; Zbl 0122.13503)] and the classical Borel-Cantelli lemma.