一种基于置信规则的模糊推理算法

由于传统模糊推理不考虑前提属性相对于结论属性的重要程度,也不能处理包含分布式结论的规则,偏重解决模糊性而忽略了随机性,造成专家系统性能降低。针对这一局限性,提出了一种基于置信规则的模糊推理算法。该算法以置信规则为知识表达方式,规则前件论域不受连续性限制,它将单个前件匹配度依据其相对重要程度加权求和,并把置信规则的分布式结论按照置信度加权合并,最后通过规则计算得出结论;根据基于置信规则的模糊推理方法对FuzzyCLIPS进行扩展,并在鱼雷规避系统仿真中验证该算法的有效性。

A Belief Rule-based Fuzzy Reasoning Algorithm

The traditional fuzzy reasoning ignored the importance of antecedent to the rule,and cannot handle the distributed conclusion.It resolve the ambiguity but ignoring the randomness.The limitations result in system performance degradation.In this paper,a new algorithm based on belief rules is proposed.The algorithm is based on the knowledge expression of belief rules,which are not be enslaved to the continuity of antecedent.It weights the sum of these single antecedent match degrees according to their relative importance degree.The distributed conclusions are combined on the basis of their degree of belief.Finally,the algorithm gets the inference on the grounds of rule calculation.The new algorithm is realized as a new part of FuzzyCLIPS.The algorithm is implemented and certificated in the torpedo evasion system simulation.