论文部分内容阅读
粒子间信息的共享方式对粒子群优化算法的收敛速度和全局搜索能力有重要的影响.针对全互联、环形拓扑结构,提出基于双层子群的信息共享方式,以收敛率作为子群规模变化的标识,实现子群规模动态变化,协调了算法的全局搜索能力和局部寻优能力.子群排斥机制使子群跳出局部最优解的束缚,提高解的多样性.选取目前比较流行的几种粒子群优化算法,通过五种经典的Benchmark高维函数优化问题进行实验仿真.结果表明基于双层可变子群的动态粒子群优化算法可以有效的避免算法陷入局部最优,在保证收敛速度的同时算法的全局搜索能力和精度有明显的提高.
The way of information sharing among particles plays an important role in the convergence speed and global search ability of particle swarm optimization algorithm.For the full interconnection and ring topology, a two-layer subgroup-based information sharing method is proposed, and the convergence rate is taken as the subgroup size change To realize the dynamic change of subgroup size and coordinate the global search ability and the local optimization ability of subgroups.Exact mechanism of subgroup exclusion makes the subgroup jump out of the shackles of the local optimal solution and improve the diversity of solution.Select the most popular Particle Swarm Optimization (PSO) algorithm, which is simulated by five classical Benchmark high-dimensional function optimization problems. The results show that the dynamic particle swarm optimization algorithm based on two-layer variable subgroup can effectively avoid the algorithm from falling into the local optimum, While the global search capability and accuracy of the algorithm have been significantly improved.