Bin zhangl, Shenglong on^', JingCheng wang', Jianrnin zhang2, HuiHe ~ h a o ' 1. Department of Automation, Shanghai Jiao Tong University 2. Automation Institute, Technology Center of Baosteel Corp. Abstract: In this paper, an approach is introduced to get complete rulebase fiom recorded data. At the beginning, a small number of fuzzy sets are defined on each variable to get a complete rulebase. In this case, even if there are missing rules, the rulebase can be completed by knowledge easily. Then, the number of fuzzy sets increased step by step until the performance of the model reach a certain level or cannot be improved further. During this procedure, if the generated rulebase is incomplete, the model constructed at last step can be used to complete the rulebase at this step. Finally, the proposed method is used to build the model of a reheating furnace and the simulation results show the effectiveness of the proposed method.
[1]
Junhong Nie,et al.
Constructing fuzzy model by self-organizing counterpropagation network
,
1995,
IEEE Trans. Syst. Man Cybern..
[2]
Shigeo Abe,et al.
Fuzzy rules extraction directly from numerical data for function approximation
,
1995,
IEEE Trans. Syst. Man Cybern..
[3]
S. Abe,et al.
A classifier using fuzzy rules extracted directly from numerical data
,
1993,
[Proceedings 1993] Second IEEE International Conference on Fuzzy Systems.
[4]
Michio Sugeno,et al.
Fuzzy identification of systems and its applications to modeling and control
,
1985,
IEEE Transactions on Systems, Man, and Cybernetics.
[5]
Li-Xin Wang,et al.
Adaptive fuzzy systems and control - design and stability analysis
,
1994
.