Research On Fuzzy Interpolative Reasoning Methods Based On Multiple Fuzzy Reasoning Model

Fuzzy reasoning is the main research about fuzzy logic, it has been widespread concerned。Many fuzzy reasoning methods have been brought forward so far,but for the merit and inferior,there is no generally accepted evaluation criteria。Via rough induction,we can see that consistency of inference.monotonicity on linguistic mood.interposal property.normality and computational simplicity should be a basic requirement。Here, monotonicity on linguistic mood is the rule that "if A then B" implies "very A then very B"。Interposal property is that when a premise is between two other,the corresponding reasoning results should be between the other two。Construct a new reasoning algorithm so that it has the nature of the fuzzy reasoning has been the main theoretical issue。To avoid various logic consideration interrelated literature hold the problem of fuzzy reasoning return to the expansion of the function, a fuzzy rule is treated as a couple of two fuzzy numbers,and a fuzzy rule-base a fuzzy interpolative nodes set。A method to construct an inference function is presented, and then the reasoning conclusion is obtained by extensase is an ordered rule set which is overlapping complement,the method cion principle。The inference is facilitated to a composite function。If rule-ban guarantee consistency oinference, monotonicity on linguistic mood, interposal property and normality,no matter that the rule-base is sparse and complete。In this paper,a fuzzy interpolative reasoning method which the antecedent of fuzzy rules is two variables is gived based on the use of bilinear interpolative function and fuzzy extension principle, the method can guarantee consistency of inference, monotonicity on linguistic mood, interposal property and normality,and the simulation experiments is done,the rusult shows that the inference algorithm is not only easy to compute but also has a lot of good nature,it is to meet the usual requiment of reasonable inference.