约束多目标人工蜂群算法

为提高约束多目标进化算法的收敛性和解集分布性,提出一种基于人工蜂群算法的改进约束多目标进化算法CMABC。在利用外部种群分别存储较优可行解和不可行解处理约束条件的基础上,根据约束多目标问题的特点,对外部种群的更新方式、迭代种群的更新方式及人工蜂群算法进行改进。实验仿真结果表明,CMABC相对于目前性能较好的MOABC及HPSO具有一定优势,能够在保证良好收敛性的同时,使获得的Pareto最优解集具有更均匀的分布性和更广的覆盖范围,适合于约束多目标优化问题的求解。

Constraint multi-objective evolutionary algorithm based on artificial bee colony algorithm

To improve the convergence and diversity of constraint multi-objective evolutionary algorithms,a novel constraintd multi-objective evolutionary algorithm,named CMABC,based on Artificial Bee Colony(ABC) algorithm is proposed.CMABC treats constraint conditions using external population storing feasible and unfeasible solutions.In addition,according to the feature of constraint multi-objective optimal problems,the update method of external and iterative population,and the ABC algorithm are improved.Experiment results show that the CMABC outperforms the state-of-the-art Multi-Objective Artificial Bee Colony(MOABC) algorithm and Hybrid Particle Swarm Optimization(HPSO) algorithm in terms of convergence and diversity matrics,which proves the superiority of CMABC in solving constraint multi-objective optimal problems.