摘要
提出一种直接优化求解控制量的轨迹跟踪控制策略，无须知道系统的先验知识，基于神经网络在线实时辨识，反复迭代求解已知跟踪轨迹的一步最优控制量。几组非线性系统的仿真表明该算法具有简捷快速，适应面宽的特点。
In this paper ,the characteristics and difficulties of scheduling problems in cell controller are analyzed, and a review of the current research is presented. On this basis, a feature oriented dynamic scheduling framework based on simulation and learning is proposed, which is used as the scheduling core of intelligent cell controller in FMS.
关键词
直接优化
神经网络
非线性系统
控制策略
FMS\ \ Intelligent cell controller\ \ Scheduling\ \ Petri net\ \ Neural network
