基于模糊集贴近度的模糊推理系统摄动研究

摘要

At present,the researches on rule perturbation measurement of fuzzy reasoning systems are all based on implication relation.Un-reasonable implication operators will cause rule perturbation error.Most of fuzzy reasoning algorithms are also based on implication relation, since they relate to the operation of fuzzy relation matrix,so their computation process are relatively complex and sometimes the reasoning result is quite different to reality.In the paper we adopt a new method of measuring the rule perturbation of fuzzy reasoning system to analyse the per-turbation of fuzzy reasoning system,which no longer uses the implication operators,but measures based on close degree of fuzzy sets before and after the rule perturbation,this differs from previous methods essentially,and the reasoning result fit more the actual situation.%目前对模糊推理系统规则摄动度量都是基于蕴涵关系进行研究的，而蕴涵算子选取不当必然会导致规则摄动出现误差。多数模糊推理算法也是基于蕴涵关系，这些算法因涉及模糊关系矩阵运算而使算法的计算过程比较复杂，有时推理结果和实际相差甚远。采用一种度量模糊推理系统规则摄动的新方法分析模糊推理系统的摄动，该方法不再使用蕴涵算子，而是基于规则摄动前后模糊集合的贴近度来度量，与以往的方法有着本质的不同，推理结果更贴合实际情况。