基于粒子群的BP神经网络算法在猪等级评定中的应用  

Study on Pork Grade Evaluation of BP Neural Network Based on PSO

作　　者：王越[1] 曾晶[1] 董丽梅[1] 张权[2] 

机构地区：[1]重庆理工大学,重庆400054 [2]西南交通大学交通运输与物流学院,四川南充637900

出　　处：《重庆理工大学学报（自然科学）》2013年第1期37-41,共5页Journal of Chongqing University of Technology：Natural Science

基　　金：重庆市科技攻关项目(2010GGB097)

摘　　要：现有的在猪等级评定中应用的BP神经网络算法存在对初始权值敏感、易陷入局部最小值等缺陷,从而导致预测精度不高、收敛速度慢的状况。针对该问题,在神经网络训练中引入基于全局随机优化思想的粒子群优化(PSO)算法。先利用PSO优化BP神经网络的初始权值,然后采用神经网络完成给定精度的学习,建立了粒子群-BP神经网络模型。与传统BP神经网络相比,该方法预测精度高、收敛速度快,可以有效地运用到猪等级评定中。The BP neural network algorithm has been widely applied to the pork grade evaluation. But because of the defect in BP neural network of pork grade evaluation at present which is sensitive with the initial weights, easy to fall into the local least value, low forecast precision and slow convergence speed occurred. This paper introduced the particle swarm optimization （PSO） algorithm based on the random global optimization into the neural network training. First the PSO is used to weights of BP neural network, and then neural network is used for given accuracy to found the PSO-BP neural network model. Compared with the traditional BP neural network, the PSO-BP optimize neural network model has the merits of faster convergence speed and higher forecast precision, and it can be effectively ap- plied to the pork grade evaluation.

关 键 词：猪等级评定 粒子群算法 BP神经网络 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]