摘要
本文提出了一种少控制参数的约束多目标微粒群优化算法.该算法利用关于微粒全局和个体最优点的高斯分布来更新微粒的位置,无需设置惯性权重和学习因子等控制参数;利用非可行储备集保存所得非可行解,给出一种改进的储备集更新方法;为均衡微粒对未知可行域和已知可行域的开发/探索能力,提出一种线性递减策略,用来分配微粒从非可行储备集中选择全局最优点的概率.最后,实验验证了所提算法的有效性.
This paper presents a constrained multi-objective particle swarm optimization algorithm with few control parameters to solve constrained multi-objective optimization problems.In this algorithm,a Gaussian distribution based on the global/local best positions is developed to update the particles＇ positions.It makes unnecessary to perform fine tuning on such control parameters as inertia weight and acceleration coefficients.Using an infeasible archive to save infeasible solutions,an improved update method of the infeasible archive is proposed.In order to balance the algorithm＇s capabilities to exploit known feasible regions and to explore unknown feasible regions,a linear decreasing strategy is introduced to assign the probability,based on which the particles select their global best positions from the infeasible archive.Finally,feasibility of the proposed algorithm is validated by simulation results.
出处
《电子学报》
EI
CAS
CSCD
北大核心
2011年第6期1436-1440,共5页
Acta Electronica Sinica
基金
国家自然科学基金资助(No.61005089)
江苏省自然科学基金资助(No.BK2008125)
高等学校博士学科点专项科研基金资助课题(No.20100095120016)
关键词
多目标优化
约束
微粒群
高斯分布
multi-objective optimization
constraint
particle swarm optimization
Gaussian distribution
