摘要
探讨了运用遗传算法,对润滑油添加剂分析系统中神经网络结构的设计,对不同结构的网络进行编码,每个码链对应一个网络结构,利用多种不同的初始连接权值分别对每个网络进行训练,利用误差函数确定适应度函数,利用交叉、变异等遗传操作算子对群体进行处理,从而得出最佳网络结构。
The neural network structure design of the intelligence analysis system of lubricate oil was studied based on the genetic algorithm. The different code was given to the different network structure, and each code series according with one network structure. Each network was trained by the various initial connection weights,fitness function was determined by error function and the group was treated by crossover and mutation operators. From this the best network structure was achieved.
出处
《润滑与密封》
EI
CAS
CSCD
北大核心
2006年第6期127-129,132,共4页
Lubrication Engineering
关键词
遗传算法
润滑油
添加剂
神经网络
genetic algorithm
lubricate oil
additive
neural network