基于局部协同与竞争变异的动态多种群粒子群算法

针对粒子群算法在处理复杂优化问题时,出现多样性较差、收敛精度低等问题,提出了基于局部协同与竞争变异的动态多种群粒子群算法(Dynamic Multi-population Particle Swarm Optimization Based on Local Cooperative and Competitive M utation,LC-DM PPSO).LC-DM PPSO算法设计了一种局部协同的方法,该方法划分种群成多个子种群,划分后的子种群再通过非支配排序、差分变异的方法选择出一对领导粒子.同时,对粒子的更新方法进行改进,让各个目标优化更加均衡,增强LC-DM PPSO算法的局部搜索能力,提高收敛精度.在LC-DM PPSO算法中,为了防止出现"早熟"收敛的情况,引入竞争变异来增加种群多样性.最后,通过选择一系列标准测试函数将LC-DM PPSO算法与3种进化算法进行比较,验证所提算法的有效性.实验结果显示,所提算法的多样性和收敛性比其他3种进化算法更好,优化效果更佳.

Dynamic Multi-population Particle Swarm Optimization Based on Local Cooperative and Competitive Mutation

In view of the poor diversity and low convergence accuracy of particle swarm algorithm when dealing with complex optimization problems, a dynamic multi-population particle swarm optimization based on local coordination and competitive mutation (Dynamic Multi-population Particle Swarm Optimization Based on Local Cooperative and Competitive Mutation, LC-DMPPSO). LC-DMPPSO designed a local coordination method, which divides the population into multiple sub-populations, and then the divided sub-populations select a pair of leader particles through the method of non-dominated sorting and differential mutation. At the same time, the particle update method is improved to make the optimization of each target more balanced, enhance the local search ability of LC-DMPPSO, and improve the accuracy of convergence. In LC-DMPPSO, in order to prevent "premature" convergence, competitive mutation is introduced to increase population diversity. Finally, a series of standard test functions are selected to compare LC-DMPPSO with three evolutionary algorithms to verify the effectiveness of the proposed algorithm. The experimental results show that the diversity and convergence of the proposed algorithm are better than the other three evolutionary algorithms, and the optimization effect is better.