Dynamic lot sizing techniques: Survey and comparison
综述了动态批量定购问题的多种启发式方法，通过数值测试比较其预期成本、风险及计算时间，发现采用Groff停止规则等新方法可显著提升标准软件系统的性能。
Abstract Numerous heuristics have been proposed in the past two decades for the dynamic lot sizing problem, many of them in APICS journals. Their relative performance is explored in extensive numerical tests measuring expected costs, risks of higher than expected costs and computer time consumed. The results indicate that users of pertinent standard software systems could benefit substantially from an incorporation of more recently proposed methods, specifically Groff's (1979) stop rule and a fathoming algorithm expanding it to a look‐ahead heuristic.