基于佳点集的约束优化进化算法  

Novel Constrained Optimization Evolutionary Algorithm Based on Good Point Set

作　　者：刘慧[1] 蔡自兴[1] 王勇[1] 

机构地区：[1]中南大学信息科学与工程学院,长沙410083

出　　处：《系统仿真学报》2009年第6期1620-1623,1636,共5页Journal of System Simulation

基　　金：国家基础研究项目(A1420060159);国家自然科学基金项目(60234030;60404021)

摘　　要：结合数论中佳点集理论和多目标优化技术,提出了一种求解约束优化问题的新算法。该算法首先把约束优化问题转化为两个目标的多目标优化问题;接着结合佳点集理论重新设计了交叉算子,新的交叉算子能够生成具有代表性的子代个体以更好地搜索空间;采用BGA变异算子增加子代个体的多样性;最后根据当前子代群体的进化信息,利用联赛选择算子或Pareto优超关系选择优胜个体进入下代群体,通过4个标准测试函数验证了算法的有效性。Novel constrained optimization evolutionary algorithm based on good point set,namely COAGPN,was proposed to tackle constrained optimization problems（COPs）.Good point set in number theory and multi-objective optimization techniques were integrated into COAGPN.Firstly,COP was transformed into a bi-objective optimization problem;secondly,the crossover operator was redesigned by using the principle of good points set,so that the new crossover operator could produce a small but representative set of points as the potential offspring.After that the BGA mutation operator was applied to offspring for enhancing the diversity of the offspring population.Finally,according to the evolutionary information of the current offspring population,a tournament selection operator or Pareto dominance was used to choose the best individuals for the next population.COAGPN was tested on 4 benchmark test functions.The computational experiments show that COAGPN can converge to optimal or close-to-optimal solutions efficiently and that COAGPN outperforms or performs similarly to the other techniques referred in terms of the quality of the resulting solutions.

关 键 词：佳点集 约束优化 多目标优化 非劣个体 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TP301[自动化与计算机技术—控制科学与工程]