Cubature formulas for function spaces with moderate smoothness (Q2465307)

The authors construct simple algorithms for \(d\)-dimensional numerical integration of function classes with moderate smoothness. These classes consist of functions \(f\in L^2([0,\,1]^d)\) whose coefficients with respect to certain multiwavelet expansions decay rapidly. These algorithms are based on 1-dimensional quadrature appropriate for the integration of wavelets and on the quadrature rule of \textit{S. A. Smolyak} [Dokl. Akad. Nauk SSSR 148, 1042--1045 (1963; Zbl 0202.39901)]. The authors present upper bounds for the worst-case error of the cubature rule in terms of the number of function calls. Further they prove lower bounds showing that this method is optimal for \(d=1\) and almost optimal (up to logarithmic factors) for \(d>1\). Numerical examples are given for certain test functions in dimensions \(d=5\) and \(d=10\).