一种模糊控制器的研究

The research on an optimal fuzzy controller

根据模糊划分和模糊等价关系的定义 ,提出了 r- ε覆盖的概念 ,归纳出其实现算法 ,即问题的实质是一个关于求超图的最小覆盖问题 ,并给出了保证算法能够得到 r-ε覆盖的证明 .利用 r-ε覆盖的概念结合可加性模糊控制器的一般形式 ,在满足与所有给定的已知样本点的误差小于给定正数 ε基础上 ,使得对应模糊规则数量少 .用物元分析的办法 ,使得逼近器的精度进一步提高 .这些特点使得新模糊控制器比目前已有的模糊控制器具有更高的精度和更简捷的形式 .

Based on the definitions of the fuzzy partition and fuzzy equivalence a concept of r ε covering is proposed, and its realized algorithm is induced, i.e., the nature of the problem is to find the least cover on a supergraph. The proof which ensures that the algorithm can obtain the r ε cover is given. By using the r ε concept together with the additive general form of fuzzy controller, a new fuzz controller is setted. Under the condition that the error is less than a given positive number ε, the rule number is the new fuzzy controller is least, these features make the new fuzzy controller with higher precision and simpler form.