求解约束优化问题的佳点集多目标进化算法  

Multi-objective Evolution Algorithm of Good Point Set for Solving Restraint Optimization Problem

作　　者：裴胜玉[1] 

机构地区：[1]广西师范大学数学科学学院,广西桂林541004

出　　处：《计算机工程》2011年第24期152-154,共3页Computer Engineering

摘　　要：结合数论中的佳点集理论和多目标优化方法,提出一种求解约束优化问题的进化算法。将约束优化问题转化为多目标优化问题,引入佳点集理论,以确保所构造的个体在搜索空间内分布均匀,设计变异算子增加个体多样性,采用分群局部搜索方式,并根据Pareto非支配关系选择群体中的优势个体。实验结果表明,该算法具有较好的稳定性。A multi-objective evolution algorithm based on good point set is proposed to tackle restraint optimization problems. Good point set in number theory and multi-objective optimization methods are integrated into algorithm. Restraint optimization problem is transformed into a bi-objective optimization problem. Combined with the principle of good point set, it makes the individuals in search space distribute more evenly. The new mutation operator is applied for enhancing the diversity of the offspring population. A sub-swarm local search operator with Pareto non-dominated is used to choose the best individuals for the next pooulation. Experimental results show that the algorithm has ~ood stabilitv.

关 键 词：进化算法 佳点集 约束优化 多目标 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]