摘要
针对现有进化算法在求解传统指派问题时因取整而影响优化效果的问题,采用了一种基于AllDifferent约束的置换离散粒子群优化算法,该算法针对指派问题中各变量不能重复取值的特点,改进了算法的迭代方式,并引入了模拟退火的差解接受准则以提高优化效果,仿真算例表明改进后的算法在质量上和时间上更具有效性。
According to the limited optimization effect of the existing evolutionary algorithms on solving the traditional assignment problems, a method of replacement discrete particle swarm optimization based on AllDifferent constraint is adopted. The algorithm is aimed at the variable features of non-repeated in the assignment problems with an improved iterated mode, and in addition the worse-solution-acceptance criteria of simulated annealing algorithm is introduced to improve the optimal results. Numerical experiments show the feasibility of the improved algorithm for large-scale assignment problems.
出处
《中国管理信息化》
2009年第21期102-104,共3页
China Management Informationization
关键词
指派问题
粒子群优化算法
混合算法
大规模
Assignment Problem
Particle Swarm Optimization
Hybrid Algorithm
Large--scale
