基于改进粒子群算法的生产批量计划问题研究

Improved particle swarm optimization algorithm for dynamic lot-sizing problem

为求解基于成组单元有能力约束的生产批量计划问题,提出了一种基于二进制粒子群算法和免疫记忆机制相结合的方法,并阐明了该方法的具体实现过程。在该方法中,采用罚函数法处理约束条件,每个粒子都代表一组可用于描述具体批量计划方案的规则组合。通过对其他文献中一个仿真实例的计算和结果比较,表明该算法在寻优能力、求解速度和稳定性等方面都明显优于文献中的遗传算法。

To solve the capacitated dynamic lot-sizing problem in group technology cell, a method based on binary Particle Swarm Optimization （PSO） algorithm and immune memory mechanism was proposed and its implementation was illustrated in detail. In this method, penalty functions were used as constraints, and each particle was used to represent a group of rules set to describe concrete batching planning. Through computation of a simulation instance and result comparison, the proposed algorithm has demonstrated its higher searching efficiency and better stability than the genetic algorithms mentioned in other literatures.