摘要
在双圆盘电流变传动系统的建模过程中,通过在常用于系统识别的神经网络BP算法中加入基于生命进化原理的遗传算法改变神经算法的物理意义和性能。建模过程和结果表明,此种方法不仅能改进BP算法的效率,并且,两种算法的结合使算法本身具有了一定的物理意义,而这种意义是由电流变液的记忆功能所确定的。
In the modelingof electrorheological actuator with double driving discs, neural network algorithm′s physical significanceand performanceis expected to be realized and improved by integratinggenetic algorithmand neural algorithm. The modeling processand results show that utilizinggenetic algorithm can improve the BP algorithm′s efficiency, furthermore, the combinationof two algorithms make the algorithm itself embodyphysical significance, which is dependent on the memory characteristicof eletrorheological fluid.
出处
《机械科学与技术》
CSCD
北大核心
2005年第3期296-298,356,共4页
Mechanical Science and Technology for Aerospace Engineering
关键词
电流变传动系统
神经网络
遗传算法
Electrorheological actuator
BP network
Genetic algorithm