Probabilistic approach to analysis of information complexity of concret multivariate approximation problem

Information complexity in the worst-case setting of multivariate approximation problem of functions from reproducing kernel Hilbert space with Gaussian kernel is considered. In the paper we obtain an upper estimate of information complexity for arbitrary error threshold and parametric dimension via probabilistic methods. The main result refines the logarithmic asymptotics of Khartov and Limar and complements the estimates by Fasshauer, Hickernell, and Woźniakowski.