摘要
针对约束多目标优化问题(CMOPs)难以平衡约束条件和目标函数的不足,提出一种基于分层环境选择策略的约束多目标优化算法(CMOEA-HES).CMOEA-HES首先采用模拟二项式交叉(SBX)和差分进化(DE)算子分别产生各自的子代种群;然后通过第一层环境选择策略从两个子代种群中选出收敛性和多样性较好的个体;接着采用第二层环境选择机制在父代种群和第一层环境选择策略选出的个体中进行选择,在多样性和收敛性的基础上选出可行性较好的个体;最后将选出的个体作为下一代进化的种群.为验证CMOEA-HES的性能,将其与5种先进的约束多目标优化算法在两组典型的测试集上进行仿真计算,实验结果表明:CMOEA-HES在求解约束多目标优化问题上更具有竞争力.
Aiming at the shortcoming that constrained multi-objective optimization problems(CMOPs)were difficult in balancing the constraints and objectives,a constrained multi-objective optimization algorithm based on hierarchical environmental selection(CMOEA-HES)was proposed.In the CMOEA-HES,simulated binary crossover(SBX)and differential evolution(DE)were first adopted to produce the offspring solutions,respectively.Then,the first environmental selection mechanism was used to choose the promising solutions with better convergence and diversity from two offspring solutions.Subsequently,the second environmental selection mechanism was conducted on the parent solutions and solutions obtained from the first environmental selection,and the feasible solutions were chosen based on the diversity and convergence.Finally,the selected solutions were chosen as population for next generation.To verify the performance of CMOEA-HES,it was simulated with five state-of-the-art constrained multi-objective optimization algorithms on two typical test suite.Experimental results show that the CMOEA-HES is more competitive in solving CMOPs.
作者
张建林
曹洁
赵付青
陈作汉
ZHANG Jianlin;CAO Jie;ZHAO Fuqing;CHEN Zuohan(College of Computer and Communication,Lanzhou University of Technology,Lanzhou 730050,China;Gansu Engineering Research Center of Manufacturing Informationization,Lanzhou 730050,China)
出处
《华中科技大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2023年第5期131-136,共6页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
国家重点研发计划资助项目(2020YFB1713600)
国家自然科学基金资助项目(62063021).
关键词
约束多目标优化问题
分层环境选择
约束处理机制
进化算法
可行解
constrained multi-objective optimization problems
hierarchical environmental selection
constraint-handling technique
evolutionary algorithm
feasible solutions