Asymptotics of running maxima for \(\varphi \)-subgaussian random double arrays (Q2157380)

Let \(\{X_{k,n}, k \geq 1, n \geq 1\}\) be a double array of \(\phi\)-subgaussian random variables that are not necessarily i.i.d., defined on the same probability space. The authors study sufficient conditions on the tail distributions of \(X_{k,n}\) that guarantee the existence of a sequence \(\{a_{m,j}, m \geq 1, j \geq 1\}\) of real numbers such that \(Y_{m,j} = \max_{1\leq k \leq m, 1 \leq n \leq j} X_{k,n} - a_{m,j}\) converge to 0 almost surely as the number of random variables \(X_{k,n}\) tends to infinity. The array \(\{Y_{m,j}\}\) is split into positive part and negative part and the results are proved separately for the two. Almost sure and lim (max) convergence of random functionals of the double arrays are investigated. After discussing some examples, some numerical examples are provided that confirm the obtained theoretical results.