摘要
本文提出了前向神经网络的分段学习算法,该算法结合自适应BP算法与 Newton 算法,从而使它既有快速收敛的特点,又克服了原算法的缺点。最后,本文将分段学习算法用于非线性自适应跟踪控制,取得了满意的效果。
A stages learning method of neural toward-networks is proposed in this paper. It combines with adaptive BP algorithm and Newton learning algorithm, keeping the merit of rapid convergence while overcoming the shortcomings of the original algorithms. Finally, using the stages learning algorithm in self -tuning tracking control, we obtain satisfactory result.
出处
《控制与决策》
EI
CSCD
北大核心
1993年第4期284-288,294,共6页
Control and Decision
关键词
神经网络
自适应控制
智能控制
artificial neural networks, BP networks , nonlinear adaptive control, intelligence control
