In: International Journal of Artificial Intelligence, Tools, Architectures, Languages, and Algorithms, Vol. 9, No. 4, pages 537-549. 2000.
Abstract: The paper proposes a new reasoning technique for fuzzy production systems when the given input knowledge is incomplete. Based on the fuzzy Petri net formalism, the proposed algorithm can infer all possible conclusions and their corresponding missing inputs. The most likely conclusion can also be determined based on the criteria or the minimum number of missing inputs as well as the degree of truth of the conclusion. In addition, the finiteness and computational complexity of the algorithm is investigated. As real decisions are typically made under incomplete input knowledge, this reasoning technique provides more realistic applications for fuzzy production systems.
Keywords: computational complexity, fuzzy petri nets, fuzzy productions systems, incomplete input inference.