摘要
提出一种自适应多样性保持的多目标粒子群算法(ADMMOPSO)。该算法引入多样性保持阈值(λα)来控制非劣解的分布,当多样性指标高于阈值λα时,引入一种基于网格的全局最优粒子的选择策略增加种群向真实Pareto前沿收敛的概率,并提升非劣解的多样性。通过4个测试问题和3个测试标准,并与其他算法进行比较,结果表明ADMMOPSO获得了质量较高的非劣解。
An improved multi-objective particle swarm optimizer with adaptive diversity maintenance is proposed（ADMMOPSO for short）.In ADMMOPSO,the diversity maintenance threshold value（λα） is introduced to control the non-inferior solution diversity.When the diversity maintenance value is equal or more than λα,the selection strategy of the global best performing particle based on the grid is introduced to increase the probability of convergence to the true Pareto front and improve the diversity.Through extensive comparison with other classical methods on three performance metrics in four test problems,the results show that ADDMOPSO has good performance.
出处
《济南大学学报（自然科学版）》
CAS
北大核心
2011年第3期296-300,共5页
Journal of University of Jinan(Science and Technology)
基金
山东省科技攻关项目(2009GG10001008)
贵州教育厅社科项目(0705204)
遵义师院课题(2007018
基07015
07017)
关键词
自适应
多样性保持
多目标优化
粒子群算法
adaptive
diversity maintenance
multi-objective optimization
particle swarm optimizer
