作者: Peter Winker , Kai-Tai Fang
DOI: 10.1137/S0036142995286076
关键词:
摘要: Efficient routines for multidimensional numerical integration are provided by quasi--Monte Carlo methods. These methods based on evaluating the integrand at a set of representative points area. A may be called if it shows low discrepancy. However, in dimensions higher than two and large number evaluation discrepancy becomes infeasible. The use efficient multiple-purpose heuristic threshold-accepting offers possibility to obtain least good approximations given points. This paper presents an implementation heuristic, assessment its performance some small examples, results larger sets with unknown