Backward Reasoning on Rule-Based Systems Modeled by Fuzzy Petri Nets Through Backward Tree

Abstract

The crux in rule-based systems modeled by Fuzzy Petri Nets (FPN) is to decide the sequence of transitions firing. In this problem, backward reasoning shows advantages over forward reasoning. In this paper, given goal place(s), an FPN mapped from a rule-based system is mapped further into a backward tree, which has distinct layers from the bottom to the top. The hierarchical structure of the backward tree provides the order of transitions firing. The nearer the top the transition, the earlier it fires. An innovative and efficient algorithm on backward reasoning through the backward tree with detailed descriptions on data structure is proposed in this paper.