遗传神经网络法在汽车故障诊断中的应用  

Application of the Inheriting Neural Network Method in Vehicle Failure Dia gnosis

作　　者：张蕾[1] 董恩国[1] 李泳鲜 

机构地区：[1]天津职业技术师范学院

出　　处：《汽车技术》2003年第1期36-39,共4页Automobile Technology

基　　金：天津市教委高校科技发展基金项目(01-20303)

摘　　要：在汽车故障诊断系统中,针对BP算法存在收敛速度慢、容易陷入局部极小值的问题,设计了用遗传算法的复制、交换、变异过程代替BP网络的反向传播过程的改进算法。结果表明,这种改进算法具有收敛速度快、推广性强的特点,极大提高了汽车故障诊断系统的效率和准确性。In the vehic le failure diagnosis system,aiming at disadvantages of BP arithmetic of slow conv ergence and easy to be trapped in the local minimum,the improved arithmetic tha t copy,exchange and mutation process of the inheriting arithmetic is substitute d for the backward propagation process of BP arithmetic is designed.It is shown by results that the improved arithmetic has advantages of fast convergence,eas y to be promoted,and can greatly improve efficiency and accu-rate of vehicle f ailure diagnosis systems.

关 键 词：遗传神经网络法 汽车 故障诊断 

分 类 号：U472.9[机械工程—车辆工程]