Note—Performance Bounds for Lot Sizing Heuristics
研究经典动态批量问题，推导了一类不考虑未来需求的启发式算法的最坏情况性能界限，证明其性能比下界为2，揭示了近似方法失效的条件。
This paper deals with the classical dynamic lot size problem without backlogging and capacity limitation. We derive worst case performance bounds for a class of lot sizing heuristics. When the considered methods are applied a decision whether to have a set-up or not in a certain period is taken without regarding the future demand. It is shown that 2 is a lower bound for the worst case performance ratio for such heuristics. The results illustrate circumstances under which the approximate techniques fail.