基于混合粒子群优化的多目标决策新方法

多目标优化问题中的一个关键在于合理地评判各有效解的优劣。通过引入灰色系统理论中灰色关联度的概念作为评判准则,结合粒子群优化算法进行有约束多目标规划问题的研究。提出了一种新的不可行解的保留策略,进化过程中以此策略保留适量的不可行解,有利于增强对约束边界附近可能的最优解的搜索,同时,针对粒子群优化算法的容易陷入局部最优的缺点,实现了以粒子群优化为载体的混合算法:即对全局极值邻域进一步混沌搜索寻优。仿真结果表明改进的算法对多目标决策问题是有效的。

A New Algorithm for Multi-objective Decision based on Hybrid Particle Swarm Optimization

It is unfathomed to assess the non-inferior solutions of the multi-objective optimization.The gray cognate is referred as a criterion in this restricted multi-objective optimization.A new strategy to keep some infeasible individuals in the swarm is proposed in order to enhance the capability of searching in the edge of the restriction.A hybrid chaos searching method based on PSO is adopted to ensure global convergence.The simulation shows that this improved algorithm is effective in multi-objective decision.