基于高斯-柯西混合变异的多目标粒子群算法
Multi-objective Particle Swarm Optimization Algorithm Based on Gaussian-Cauchy Mixture Mutation
舒一鸣 1戴毅茹1
作者信息
- 1. 同济大学CIMS研究中心 上海 201804
- 折叠
摘要
针对MOPSO优化算法在解决复杂的多目标优化问题上收敛表现差,搜索全局能力不足与易于陷入局部最优的缺陷,提出了一种基于高斯-柯西混合变异的多目标粒子群算法(GC-MOPSO).该算法使用一种混合高斯变异与柯西变异的变异扰动机制来提升粒子在局部与全局的搜索能力,在外部档案中采用锦标赛选择机制选取全局最优个体的策略来增加种群的多样性.通过与六项其他算法在反世代距离(IGD)上进行比较,验证了该算法的优势.
Abstract
Aiming at the defects of poor convergence performance,insufficient global search ability and easy to fall into local optimization in MOPSO optimization algorithm for solving complex multi-objective optimization problems,a multi-objective parti-cle swarm optimization algorithm based on Gaussian-Cauchy mixed mutation(GC-MOPSO)is proposed.The algorithm uses a muta-tion disturbance mechanism of mixed Gaussian mutation and Cauchy mutation to improve the local and global search ability of parti-cles,and uses the tournament selection mechanism to select the global optimal individual in the external file to increase the diversi-ty of the population.The advantages of the algorithm are verified by comparing with six other algorithms in anti-generation distance(IGD).
关键词
多目标优化/粒子群优化算法/高斯-柯西变异/锦标赛选择Key words
multi-objective optimization/particle swarm optimization algorithm/Gaussian-Cauchy variation/tournament selection引用本文复制引用
基金项目
上海市自然科学基金项目(19ZR1461500)
出版年
2024
国家科技期刊平台
NSTL
万方数据