摘要
针对BP网络存在易陷入局部极小和收敛速度慢的问题,采用遗传算法(GA)优化BP网络;并采用混配的方法,对遗传算法进行了改进,克服了遗传算法中所存在的种群内过早收敛的缺点,并在光纤陀螺温度漂移建模中,取得了较好预测的效果。
To solve the local minimum and slowly converging problem in BP neural network, genetic algorithm to optimize the BP neural network is used. A novel method named commingle copulation is used to overcome the shortcoming of converging too fast existing in species in genetic algorithm. The simulation result shows that method could be used to identify the temperature drift model of FOG.
出处
《科学技术与工程》
2008年第14期3770-3772,3782,共4页
Science Technology and Engineering
基金
国家863项目资助
关键词
人工神经网络
遗传算法
混配
光纤陀螺
温度漂移
artificial neural network genetic algorithm commingle copulation fiber optic gyroscope temperature drift
