基于混合策略的双种群约束优化算法  

Dual population constrained optimization algorithm with hybird strategy

作　　者：毕晓君[1] 张磊[1] 

机构地区：[1]哈尔滨工程大学信息与通信工程学院,哈尔滨150001

出　　处：《控制与决策》2015年第4期715-720,共6页Control and Decision

基　　金：国家自然科学基金项目(61175126);中央高校基本科研业务费专项项目(HEUCFZ1209);教育部博士点基金项目(20112304110009);辽宁省博士科研启动基金项目(201205118);辽宁省教育厅科学技术研究项目(L2012458)

摘　　要：提出一种基于混合策略的双种群约束优化算法.利用双种群存储机制处理约束条件,并采用约束支配更新不可行解集,同时采用混合策略进化种群:在进化前期利用Deb准则产生可行解,并保留一部分非劣不可行解参与进化,保持种群多样性;在进化后期让最优个体和次优个体参与进化,使种群快速收敛.仿真实验结果表明,所提出的算法在保证种群多样性的同时,能够较好地收敛到全局最优解,且鲁棒性较好.A dual population constrained optimization algorithm with the hybrid strategy is proposed. The method uses the dual population to deal with constrains, and employs constraint control as the approach of updating infeasible sets. The hybird strategy is used to evolve population. In the early stage of the evolution, the Deb criterion is used to generate feasible solutions, and some inferior infeasible solutions are retained to participate in the evolution to ensure the diversity of the population. In the later stage of the evolution, the global optimal solution and second-best solution take part in the evolution, so as to quickly converge. The simulation results show that, while keeping the diversity of the population, the proposed algorithm can converge to the global optimal solutions with better robustness.

关 键 词：约束优化 双种群 混合策略 差分进化算法 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]