Asymptotic distribution for the sum and maximum of Gaussian processes (Q2725293)

Let \( (X_{ni}) \) be a Gaussian sequence of rv's and put \( S_n = \sum_{i=1}^n X_{ni} \), \( M_n = \max _{1 \leq i \leq n} X_{ni}\). The authors investigate under which conditions on the growth of the correlation \( \sigma_n(i,j) = E X_{ni} X_{nj} \) the sum \( S_n \) and the maximum \( M_n \) are asymptotically independent if properly normalized. In the last Section 3 the results are extended to continuous time stationary Gaussian processes.