基于粒子群优化的神经网络预测模型  

Neural Network Prediction Based on Particle Swarm Optimization

作　　者：李丙春[1] 孙蕾[2] 

机构地区：[1]喀什师范学院网络中心,新疆喀什844000 [2]西安电子科技大学经济管理学院,陕西西安710071

出　　处：《新疆大学学报（自然科学版）》2007年第4期481-485,共5页Journal of Xinjiang University(Natural Science Edition)

摘　　要：传统的BP神经网络训练算法,导致训练时间长且易于陷入局部极小点.本文将粒子群优化算法用于神经网络预测模型的学习训练.实验结果表明,基于粒子群优化的神经网络学习算法更易于实现,且能更快地收敛于全局最优解.The training of conventional back-propagation method with shortcoming of spending long time and easily plunging into the local minimum. A training algorithm for neural network based on particle swarm optimization （PSO） was investigated. The results show that the training algorithm based on PSO is easily implemented and come to global optimization quickly.

关 键 词：粒子群优化算法 神经网络 全局最优 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]