A remark on the multiparameter law of the iterated logarithm (Q1123478)

Let \(B\equiv \{B(s,t)\); \(s>0\), \(t>0\}\) be a mean-zero Gaussian process with covariance function \[ E[B(s,t)B(\sigma,\tau)]=(s\wedge \sigma)\quad (t\wedge \tau) \] where \(a\wedge b\) denotes min(a,b). B is called a Brownian sheet. It is known that \[ \overline{\lim}_{T\to \infty}\sup_{st=T,\quad s,t\leq T}B(s,t)/(4T \log \log T)^{1/2}=1\quad a.s. \] [cf. \textit{G. J. Zimmerman}, Ann. Math. Statistics 43, 1235-1246 (1972; Zbl 0244.60032)]. Here the author provs that \[ \underline{\lim}_{T\to \infty}\sup_{1\leq s\leq T,\quad st=T}B(s,t)/(2T \log \log T)^{1/2}=1\quad a.s. \] Generalizations of this result are discussed.