一种基于佳点集原理的约束优化进化算法  

Constrained optimization evolutionary algorithm based on good lattice points principle

作　　者：肖赤心[1,2] 蔡自兴[1] 王勇[1] 周经野[2] 

机构地区：[1]中南大学信息科学与工程学院,长沙410083 [2]湘潭大学信息工程学院,湖南湘潭411105

出　　处：《控制与决策》2009年第2期249-253,258,共6页Control and Decision

基　　金：国家自然科学基金项目(60234030,60404021);国家基础研究项目(A1420060159);湖南省院士基金项目(06IJY3035)

摘　　要：提出一种基于佳点集理论解决约束优化问题的进化算法.它将实分圆域中均匀分布的佳点映射到求解问题的搜索空间,使得所构造的个体能在搜索空间内分布比采用随机方式更加均匀,并引进预交叉机制来平衡佳点取点个数与算法搜索能力之间的矛盾.新算法的遗传算子基于佳点技术构造,精度不受空间维数的限制,有利于高维优化问题.对6个标准测试函数的数值实验结果验证了新算法的通用性、有效性和稳健性.An approach is proposed to handle constrained optimization problems by using evolutionary algorithm. It maps the good lattice points （GLP） from S-dimensional unit hypercube to the S-dimensionaI search space, which makes the individuals in search space distribute more evenly than the random methods do. In order to balance the quantity of GLP and the exploitive capability of proposed algorithm, a new pre-crossover operator is proposed. The gene operators of the new method are constructed according to GLP principle whose precision is not be confined to the dimension of the search space, which is benefit for some high dimensional optimization problems. The new approach is compared with other evolutionary optimization techniques by six benchmark functions. The results obtained show that the new approach is general, effective and robust.

关 键 词：进化算法 佳点集 约束处理技术 约束优化 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]