基于混沌优化和网格筛选策略的多目标分布估计算法

为了解决多目标分布估计算法中进化速度慢、解精度和分布不佳等问题,提出一种基于混沌优化和网格筛选策略的多目标分布估计算法.该算法首先利用混沌模型进行种群的初始化,以获得较理想的初始化结果;然后运用混沌的局部优化策略对每代产生的非支配个体进行寻优,加速种群向Pareto最优前沿的逼近;最后利用简单的网格筛选策略保持个体的均匀分布,从而增强精英种群的多样性.3种评价标准在8个测试问题上的实验表明:与目前最具代表性的RM-MEDA算法相比,该算法不仅在接近真实的最优前沿和保持种群的多样性方面具有一定优势,而且在进化速度上也有较大提高.

Multi-objective Estimation of Distribution Algorithm Based on Chaos Optimization and Grid Selection

To solve the poor performances of evolution speed,solution precision and distribution in the multi-objective estimation of distribution algorithm,this paper proposes a new algorithm that based on chaos optimization and grid selection strategies. The algorithm first performs initialization using chaos models to obtain better initial results. Then,a chaotic local optimization strategy is applied to get nondominating individuals in iterations,which makes the population effectively approximate the Pareto optimal front. Finally,a simple grid selection strategy is employed to keep a uniform distribution and enhance the diversity of the elite population. Experimental results on eight test problems using three performance metrics show that the new algorithm has a certain advantage compared to the most representative RM-MEDA algorithm in terms of converging to the true Pareto front and maintaining the diversity of the population,moreover,it is also much faster than RM-MEDA.