带有个体扰动和相互学习改进的粒子群优化算法

Improved Particle Swarm Optimization Algorithm with Individual Disturbance and Mutual Learning

针对粒子群优化算法容易陷入局部极值、进化后期早熟收敛现象,提出了一种带有个体扰动和相互学习改进的粒子群优化算法.算法在迭代的过程中,根据群体适应度方差按照一定的概率对当前的个体最优粒子进行扰动,增强了算法的局部探索的能力,使得粒子跳出局部最优点;同时增加粒子的相互学习阶段,使得每个粒子的进化不仅受到个体最优粒子和全局最优粒子的影响,而且还受到其他粒子之间相互学习的影响,提高了算法的收敛速度.数值实验表明,改进的新算法具有更高的收敛速度和收敛精度,能有效克服早熟收敛现象.

In order to overcome the shortcomings of the particle swarm optimization （PSO） algorithm, such as easily falling into local extremum and the premature convergence phenomenon at the later evolution process, we propose an improved PSO algorithm with individual disturbance and mutual learning in this paper. In every iteration, the current optimal individual is disturbed by the group fitness variance with a certain probability, which enhances the ability of local exploration and makes the particles jump out of local optimal points. Meanwhile, we add the mutual learning phase, which makes each particle evolution influenced not only by the personal best and global best, but also affected by mutual learning between particles, and improves the convergence speed of the algorithm. Numerical experiments show that the improved algorithm has higher convergence speed and accuracy. It also can effeetively conquer the premature convergence phenomenon.