论文部分内容阅读
分析高斯动态粒子群优化算法(GDPSO)中新的种群产生方式的特点,针对传统粒子群优化算法中全局最优模型收敛速度快但易陷入局部最优、局部最优模型收敛速度较慢的缺点,提出一种新的粒子群信息共享方式——多簇结构.该算法在簇内部实现粒子间信息的高度共享,而在簇之间则通过松散的连接实现信息的传递,以协调 GDPSO 算法的勘探和开采能力.通过典型的 Benchmark 函数优化问题测试并分析经典拓扑以及多簇结构在GDPSO 算法中的性能,仿真实验结果表明,采用特定多簇结构的 GDPSO 算法收敛速度和稳定性显著提高,同时全局搜索能力明显增强.
In this paper, the characteristics of new population generation methods in Gaussian dynamic particle swarm optimization (GDPSO) are analyzed. In order to overcome the shortcomings of the traditional global optimal particle swarm optimization algorithm such as fast convergence, easy convergence in local optima and slow convergence in local optimal models , A new multi-cluster information sharing scheme based on particle swarm optimization (PSO) is proposed, which is to share information among particles in a cluster and transmit information through loose connection between clusters to coordinate the GDPSO algorithm Exploration and exploitation ability.Through the typical Benchmark function optimization problem testing and analyzing the performance of classical topology and multi-cluster structure in GDPSO algorithm, the simulation results show that the convergence speed and stability of GDPSO algorithm with a particular multi-cluster structure are significantly improved, meanwhile, Global search capabilities significantly enhanced.