一种新的基于模拟退火的粒子群算法

A New Particle Swarm Optimization Algorithm Based on Simulated Annealing

鉴于标准粒子群优化算法易陷入局部最优、收敛精度低,我们提出了一种改进的基于模拟退火的粒子群算法(NPSO)。将模拟退火算法的思想引入粒子群算法中,并对更新公式进行简化;提出了一种自适应随机惯性权重,实现了自适应平衡局部搜索和全局搜索的能力;提出了"优胜劣汰"的更新机制,加快了算法的收敛速度。与其它几种粒子群算法在4个基准测试函数上的实验比较,实验研究表明,NPSO算法的性能很好。

Because the standard particle swarm optimization algorithm is easy to fall into local optima and conver-gence accuracy is low, we propose an improved particle swarm optimization based on simulated annealing (NPSO). The idea of simulated annealing algorithm is involved into particle swarm optimization and update formula is simplified. To realize self-adaptive balance local search capability and global search capability, we propose a self-adaptive random inertia weight. “Survival of the fittest” update mechanism is proposed to accelerate the convergence rate. The experi-mental study shows that NPSO has a better performance in comparison with several variant PSO algorithms on four benchmark functions.