求解多目标优化问题的自适应混沌混合蛙跳算法

ADAPTIVE CHAOS SHUFFLED FROG LEAPING ALGORITHM FOR MULTIOBJECTIVE OPTIMISATION SOLUTION

针对多目标优化问题提出一种自适应混沌混合蛙跳算法MACSFLA(Adaptive chaos shuffled frog leaping algorithm for multiobjective optimization)。使用动态权重因子策略以提高混合蛙跳算法SFLA(Shuffled Frog Leaping Algorithm)收敛效率,引入基于Pareto支配能力的SFLA子族群划分策略,使得SFLA能够应用于多目标优化问题。在此基础上,MACSFLA首先利用SFLA快速寻优能力接近理论Pareto最优解,然后采用自适应网格密度机制动态维护外部存储器Pareto最优解规模,并使用自适应混沌优化技术改善Pareto最优解集样本多样性,最后利用Pareto最优解选择策略为青蛙种群选择最优更新粒子。多目标函数测试实验结果表明,与MOPSO和NSGA-Ⅱ相比,MACSFLA在Pareto最优解集均匀性和多样性上有明显优势。

We propose an adaptive chaos shuffled frog leaping algorithm （MACSFLA）for multi-objective optimisation problem.It uses dy-namic weighting factor strategy to improve the convergence efficiency of shuffled frog leaping algorithm （SFLA）,and introduces Pareto control capability-based SFLA sub-ethnic partition strategy to make SFLA be able to apply to multi-objective optimisation.On this basis,MACSFLA first employs fast search ability of SFLA to approach the optimal solutions of theoretical Pareto,and then uses adaptive grid density mechanism to dynamically maintain the scale of optimal Pareto solution of external memoriser,and uses adaptive chaos optimisation technology to improve the sample diversity of optimal Pareto solution.Finally,it uses optimal Pareto solution selection strategy to select more update particles for frog populations.Results of multi-objective function test experiment show that,compared with MOPSO and NSGA-Ⅱ,MACSFLA has evident ad-vantages in uniformity and diversity of optimal Pareto solution set.