摘要
针对过程控制中工程化建模方法 ,得到的对象模型为传递函数形式 ,而目前的研究是化为状态空间表示 ,因而迭代学习算法及收敛性条件是以矩阵的形式表示 ,不太适合于实际应用的事实。文章用时频结合的方法给出了设定值变化下的迭代学习算法 ,并以系统参数形式地给出几种学习算法的收敛性定理 。
As to the method of engineering modeling in process control, the object model here obtained is represented by transfer function. However, the current research is to transfer it as the form of state space, hence, iterative learning algorithm and its convergence conditions are represented by matrix, which doesn't match the practical application. Here, time-frequency combination method is applied to obtain the iterative algorithm with the assumed number changing. The convergence theorems for the learning algorithms based on the system parameters are also obtained.In the meantime,the empirical simulation supports the results.
出处
《中山大学学报（自然科学版）》
CAS
CSCD
北大核心
2002年第2期15-18,共4页
Acta Scientiarum Naturalium Universitatis Sunyatseni
