粒子群算法中随机数参数的设置与实验分析

粒子群算法的相关参数,对粒子群算法的优化性能有着重要影响,本文针对粒子群算法模型中随机数参数的设置问题展开实验分析.首先,由于各种高级程序语言的结构不同,在粒子群算法的实现程序中,对速度更新公式内同一个粒子速度向量,其各个分量的随机数参数的设置各不相同.其次,根据连续函数优化问题和作业车间调度问题中的典型测试算例,以及对于设备拥有量参数优化问题的计算,表明在粒子群算法中设置不同的随机数参数将对粒子群算法的优化性能产生较大影响,并且,对一个粒子速度向量中的不同分量所对应的随机数参数,如果设置相同的值,可以有效地提高粒子群算法的优化效率.

Parameter setting and experimental analysis of the random number in particle swarm optimization algorithm

The parameters in particle swarm optimization have important effect on the optimization performance. The parameter setting of the random number in the particle swarm optimization model is analyzed by the experiments. First, because of different structures in different high-level languages, we find that in the program of particle swarm optimization algorithm, different components of a velocity vector may have different parameter settings for the corresponding random number in the particle velocity updating equation. Next, in continuous function optimization and benchmark tests of Job Shop scheduling, as well as the computation of the equipment-possession-quantity parameter optimization model, all results indicate that different parameter settings for the random number may cause significantly different effects on the optimization performance of particle swarm optimization algorithm. Furthermore, it is also found that the optimization efficiency of a particle swarm optimization algorithm can be obviously improved if the corresponding random number of different components of a velocity vector is set to the same value.