An operator theoretical characterization of \(\epsilon\)-entropy in Gaussian processes (Q1074554)

The purpose of this paper is to study a relation between two kinds of \(\epsilon\)-entropies, that is, one is of compact operators and another is of stochastic processes. Let \({\mathcal H}\) be a Hilbert space, T be a compact operator on \({\mathcal H}\) and \(\xi =\{\xi (t)\); \(0\leq t\leq 1\}\) be a mean continuous stochastic process. Denote by S(T,\(\epsilon)\) and H(\(\xi\),\(\epsilon)\) the \(\epsilon\)-entropy of T and that of \(\xi\) respectively. Then the asymptotic behavior of the sequence \(\{S(T^ k,\epsilon^ k)\); \(k\in {\mathfrak N}\}\) are discussed and, by using this result, the relation between the \(\epsilon\)-entropy of \(\xi\) and that of the integral kernel operator which is generated by the covariance function of \(\xi\) are investigated. This relation is applied to the estimation of the order of growth of H(\(\xi\),\(\epsilon)\).