Complexity of stochastic integration in Sobolev classes (Q2633847)

This is an advanced study in information-based complexity. The objective is to study stochastic integration \(S(f) := \int_{\Omega} f(x)\,d\hat W_{Q}(x)\) on some bounded Lipschitz domain \(\Omega\) for deterministic integrands \(f\) and with respect to an isonormal Gaussian process \(\hat W_{Q}\). In particular, due to the randomness in the stochastic integral the corresponding solution operator \(S\) is random. The first four sections are devoted to thoroughly introducing notation and auxiliary results (sometimes going beyond existing theory) on lower and upper bounds for general numerical problems. The main complexity results for stochastic integration are given in Section~6. It is shown that the complexity for this problem is closely related to embedding in Sobolev spaces. Therefore, the upper bounds depend on results previously obtained in [\textit{S. Heinrich}, J. Complexity 25, No. 5, 455--472 (2009; Zbl 1173.65304)].