PSO优化的LS-SVM在列车弓网系统的建模研究

Modeling of Train Pantograph-catenary System Based on LS-SVM Optimized by PSO

提出基于粒子群(PSO)优化最小二乘支持向量机(LS-SVM)的列车弓网系统建模方法。针对LS-SVM的超参数难以选择的问题,提出采用具有全局搜索性能的PSO优化LS-SVM超参数的方法。在建立弓网子系统模型的基础上,得到了弓网系统的整体动力学方程。最后进行弓网系统的仿真实验,结果表明,所提出的PSO优化LS-SVM模型比LS-SVM模型、子空间模型具有更高的预报精度,所提出的方法用于列车弓网系统的建模是有效的。

A modeling method, based on least squares support vector machine （LS-SVM） optimized by particle swarm optimization （PSO）, for train pantograph-catenary system is proposed in this paper. For the difficult choice of hyper parameters of LS-SVM, the PSO with global search performance is used to optimize the hyper parameters of LS-SVM. In addition, the whole pantograph-catenary dynamic equation is obtained on the basis of establishing the model of the pantograph-catenary subsystem. Finally, simulation experiment is carried out, and the simulation results show that proposed LS-SVM optimized by PSO model has better forecast precision than LS-SVM model.Tthe proposed method is effective in modeling train pantograph-catenary system