Good lattice rules in weighted Korobov spaces with general weights (Q2491143)

Quasi Monte-Carlo algorithms are considered for multivariate integration. The worst-case error of the algorithms is defined in terms of weighted Korobov space norm with arbitrary weights (not necessarily a tensor product of univariate functions spaces). The existence of good lattice rules and sufficient conditions of tractability or strong tractability in such weighted spaces are demonstrated. A component-by-component algorithm for constructing good lattice rank-1 rules is proposed. This rule have optimal order of convergence. Numerical results for the problem of path-dependent options pricing are described.