基于动态改变惯性权值的粒子群算法

粒子群算法的速度更新公式是通过惯性权值来调节前代速度对当代速度的影响。标准粒子群算法的惯性权值是采用线性递减策略,这使得算法极易收敛到局部最优;而且这种方法依赖于最大迭代次数的设定,使得惯性权值的选取具有盲目性。本文提出一种动态改变惯性权值的方法,充分利用目标函数所提供的信息,构造按指数衰减的惯性权值并进行了分析,最后对一标准测试函数进行了仿真。结果表明,所提算法能够得到更好的优化效果,验证了方法的有效性。

Study of particle swarm optimizer algorithm based on dynamic inertia weight

The influence level of the previous generation speed to contemporary speed is adjusted by the inertia weight in the formula of speed updating in particle swarm optimization algorithm.The inertia weight of standard PSO adopts linear reduction strategy,which makes the algorithm constringe the local optimum easily~ furthermore,this approach depends on the number of the largest iteration making the choosing of inertia weight be in blindness.So the way of changing iner- tia weight dynamic is presented in the paper with the full use of the information provided by the objective function,and the inertia weight with index attenuation is constructed and analyzed,and finally a standard testing function is simulated by this method.The simulation results show that the modified algorithm can get better optimization effect,which vali- dates the effectiveness of the method.