基于命题空间矩阵变换的模糊Petri网推理

Fuzzy Petri Net Reasoning Based On Proposition Space Matrix Transformation

针对模糊petri网推理算法存在并行推理效率低下,且难以获得命题库所间的函数关系的问题,提出一种命题空间变换的模糊petri网推理方法,从系统整体出发,将所有命题与变迁构建命题空间矩阵;以命题库所的状态为向量联立初始命题库所的空间矩阵,得出命题库所状态下的常数向量,并以析取式的模糊petri网规则择取相应命题空间矩阵,联立常数向量并行求解最终的命题库所状态。通过推理算法的对比分析,改进算法在并行推理效率上得到了提高,并获得命题库所间的函数关系,对揭示命题库所间的内在联系和模糊petri网的参数学习提供了一个新的方法。

This paper proposed a fuzzy Petri net reasoning method based on propositional space transformation. Based on the whole system, this method constructed the proposition space matrix of all propositions place and transi- tions. The constant vector of the proposition place state was obtained by the state vector of the proposition place and the initial proposition place of space matrix. Disjunctive FPR choosed the corresponding proposition space matrix and contacted the constant vector parallel solving the final proposition database states. By comparing the inference algo- rithm analysis, the parallel inference efficiency of this algorithm has been improved, and which obtains the function relationships between proposition places, provides a new direction in order to further reveal the inner link of the prop- osition place and studies the parameters of fuzzy Petri nets.