CIMS下MRP-Ⅱ多级生产批量计划问题的研究

Research for Lot-Sizing Problem of MRP-Ⅱ in CIMS

多级生产批量计划问题的研究与开发不但是一个应用上的难题，也是一个理论的难题[1]．本文基于问题的性质．从一个新的角度即从调整变量出发，运用遗传算法（GA）随机搜索进行求解；通过仿真实验，测试6个问题表明，多级生产批量计划问题的GA算法比循环动态规划启发式算法平均改善5％以上．

The multi-level lot sizing problem is not only theoreticall 'hard', but also computationally very difficult handle. In this paper we propose a new genetic stochastic search method to solve MLLS based on properties of the solution of MLLS; Computational results of 6 tested problems show that the average improvenment percent given by MLLS/GA algorithm is over 5 % as compared to MLLS/Heuristic solutions.