对模糊推理算法的一点思考

Some Considerations of Fuzzy Inference Algorithm

模糊推理是模糊控制器的主要功能,其算法的好坏直接决定控制器的效果.针对文献《智能控制基础》[1]所给一个有关模糊推理的结论,笔者经过实例检验,发现该结论并不成立,于是对此问题进行深入研究,通过严格数学分析,得出正确结论,并将此结论推广到n输入模糊推理的情形.另外,将两输入Mamdani模糊推理推广到n输入模糊推理的情形.这样,文献[1-2]中的结论是本文结论的特殊情况,而且使得模糊推理计算更加简便.最后举出1个实例,计算结果验证了所得结论的正确性.

Fuzzy inference is the main function of fuzzy controller,the quality of which will determine the control performance.However,the author,through a case study,found that a conclusion about the fuzzy inference in the Literature of Fundamentals of Control Theory does not hold.Thus a further study is carried out by strict mathematical analysis.Then the correct result is obtained and extended to n-input fuzzy inference cases together with two-input Mamdani fuzzy inference.It indicates that the conclusion of the literature is just one of the special cases of the research results,and it is more simpler for fuzzy inference by using these new results.An example is also presented to illustrate the correctness of the results.