一种基于模糊Petri网的双向并行推理算法

基于模糊Petri网的并行推理算法的矩阵维数越大,其算法的时间复杂度也就越高。针对反向搜索压缩模糊Petri网模型的相关理论和并行推理算法的特点,结合矩阵命令提出一种实现双向推理的矩阵运算机制,以及其对应的基于模糊Petri网的双向并行推理算法。在使用一般模糊推理算法的过程中,推理矩阵为(11×8)维的模糊Petri网模型,而使用改进算法进行双向推理时所涉及的推理矩阵阶数仅为(7×6)。实验结果表明,与一般的模糊推理算法和反向搜索算法相比,该算法能够提高整个推理过程的并行度,降低算法的时间复杂度,从而提高推理效率。

A Bi-directional Parallel Reasoning Algorithm Based on Fuzzy Petri Nets

Time complexity of the parallel reasoning algorithm based on Fuzzy Petri Nets（FPN） is related to the dimension of matrix, and it will increase when the scale of the FPN becomes larger. By analyzing the characteristics of the parallel reasoning algorithm and the relevant theories of the Reverse Search（RS）, this paper proposes a novel Bi-directional Parallel Reasoning（BDPR） algorithm based on FPN. As for the model of FPN with the dimension of 11 rows and 8 columns, if using the BDPR algorithm, the reasoning matrix order is 7 rows and 6 columns. Experimental analysis shows that the BDPR algorithm can effectively improve the parallelism of the whole process of reasoning, reduce the time complexity of algorithm, and improve the efficiency of reasoning, compared with a general Fuzzy Reasoning（FR） algorithm and an RS algorithm.