提出并研究一种新的模糊推理算法,真域贴近模糊推理算法。给出真域贴近模糊推理算法的基本原则,利用单点模糊化方法研究真域贴近模糊推理算法与CRI方法之间的联系,给出采用最大-最小贴近度及常用蕴涵算子的真域贴近模糊推理算法,通过实验比较真域贴近模糊推理算法与CRI方法及全蕴涵三Ⅰ算法的推理能力。理论分析及数值实验表明,真域贴近模糊推理算法不仅具有较好的语义逻辑基础,而且具有较好的推理能力。

A novel fuzzy reasoning method named fuzzy reasoning method by optimizing the similarity of truth-tables(OS method for short) was proposed and discussed. Firstly,the principle of the OS method was given according to the analysis of the basic idea of the OS method. Secondly,the relationship between the OS method and CRI method was discussed. Thirdly,the OS methods with max-min similarity measure and some common implications were investigated. Finally,numerical examples were given to compare the inference abilities of the OS method,CRI method and triple I method. The theoretical analysis and numerical examples showed that the OS method had a good logical foundation and good inference ability.