求解生产批量计划问题的改进量子进化算法  

Improved quantum-inspired evolutionary algorithm for lot size scheduling problem

作　　者：孙棣华[1] 解佳[1] 赵敏[1] 

机构地区：[1]重庆大学自动化学院,重庆400044

出　　处：《计算机集成制造系统》2010年第8期1702-1707,共6页Computer Integrated Manufacturing Systems

基　　金：国家863计划资助项目(2006AA04A124)~~

摘　　要：针对遗传算法和粒子群算法在求解生产批量计划问题中易陷入局部最优解的问题,提出了改进的量子进化算法。对各周期项目计划产量的决策变量进行基于概率幅的量子比特个体编码,在迭代求解的过程中通过约束违反度比较个体的支配关系,有效指导种群向合理解进化,并根据当前迭代次数动态调整旋转角机制控制基因位的坍塌速度,在进化后期尽量保留最优个体的基因信息以提高算法的收敛速度和求解精度。实验结果表明了该算法的有效性。An improved quantum-inspired evolutionary algorithm was proposed to solve lot size scheduling problem,against Genetic Algorithm（GA）and Particle Swarm Optimization（PSO）algorithm falling in the convergence to the local optimal solution.Decision variables of the product planning output in each period were encoded with amplitude amplifications.During iterative solution process,dominance relationships of individuals were compared to the con-straint violation degree,which guided population convergence to the reasonable solution.Meanwhile the quantum collapse was controlled by the mechanism of dynamically adjusting rotation angle according to current iteration times.In the later phase of evolution,to speed up convergence,genetic information of the optimal individual was saved as much as possible.Experimental results verified the effectiveness of this algorithm.

关 键 词：生产批量计划 量子进化算法 旋转角 约束违反度 整数规划 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]