带变异算子的自适应粒子群优化算法

Adaptive particle swarm optimization algorithm with hybrid mutation operator

针对粒子群优化算法在进化过程的后期收敛速度较慢,易陷入局部最优的缺点,对基本粒子群优化算法作了如下改进:在速度更新公式中引入非线性递减的惯性权重;改进位置更新公式;对全局极值进行自适应的变异操作。提出一种新的混合变异算子的自适应粒子群优化算法。通过与其他算法的数值实验对比,表明了该算法具有较快的收敛速度和较好的收敛精度。

A modified Particle Swarm Optimization（PSO） is proposed to improve the performance of standard PSO that is easily trapped in local optimum and has a slow convergence rate in the late period.On the basis of standard PSO,the modified algorithm applies some methods such as citing a nonlinearly descending inertia, changing the velocity iteration formula and introducing the mutation operator during the running time.The experimental results show that the new algorithm has great advantage of convergence property over standard PSO.