组合优化问题中GA性能分析及多样性评价  

Performance Analysis and Diversity Evaluation of Genetic Algorithms in Combinatorial Optimization

作　　者：武晓今[1] 韩生廉[2] 

机构地区：[1]上海交通大学电子信息与电气工程学院,上海200030 [2]同济大学电子与信息工程学院,上海200092

出　　处：《小型微型计算机系统》2005年第5期830-832,共3页Journal of Chinese Computer Systems

摘　　要：随着生产调度、机器学习、最优规划等组合优化问题的大规模化,复杂化,传统的基于运筹学的搜索算法已显得无能为力.具有广域搜索能力的遗传算法(GA)也因“完备性”与“健全性”的不充分不能有效地对应上述问题.为此,本文提出了保证GA上述两个性质地方法,使其能有效地解决复杂组合优化问题.With the large-scale and complexion of those combined optimization such as production scheduling, machine learning and optimization layout, the traditional optimization algorithm based on operational research had not be adapt for resolving those problems. While GA, which has wide scope searching capacity, also could not resolve those problems because of its insufficiency of completeness and soundness. This paper proposed a method that can guarantee the GAs two characters. It can resolve complex combined optimization problems effectively.

关 键 词：完备性 健全性 致死染色体 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]