现代电子技术2016，Vol.39Issue(11)：113-115,3.DOI:10.16652/j.issn.1004-373x.2016.11.027
自适应变系数PSO-RBF算法及其在预测工程的应用
Adaptive variable coefficient PSO-RBF algorithm and its application in forecasting engineering
林大志 1王锐利2
作者信息
- 1. 河南牧业经济学院，河南 郑州 450045
- 2. 济源职业技术学院，河南 济源 454650
- 折叠
摘要
Abstract
The RBF neural network has good effect on nonlinear prediction,but it is easy to fall into local minimum and has slow convergence. An adaptive variable coefficient PSO(particle swarm optimization)algorithm used to optimize the initial pa⁃rameters of RBF neural network is studied. And then the precise optimization for the network parameters after PSO is performed by means of RBF neural network to improve the stability,convergence efficient and precision of neural network. In adaptive variable coefficient PSO algorithm,the scheme of adaptive decreasing and increasing factors,and inertia weight factor with adaptive control is introduced into the conventional PSO algorithm to improve the algorithm ergodicity in search space,probability of finding the global optimal solution,and convergence precision and efficiency. Finally,the nonlinear relationship between gas consumption and steel output in steelmaking process is studied. The application results show that the prediction model based on adaptive variable coefficient PSO⁃RBF neural network has good prediction ability,and can play an important role in prediction engineering.关键词
非线性预测/RBF神经网络/自适应变系数粒子群算法/煤气量预测Key words
nonlinear prediction/RBF neural network/adaptive variable coefficient PSO algorithm/gas volume prediction分类
信息技术与安全科学引用本文复制引用
林大志,王锐利..自适应变系数PSO-RBF算法及其在预测工程的应用[J].现代电子技术,2016,39(11):113-115,3.
