Reinforcement learning for high-level fuzzy Petri nets.

The author has developed a reinforcement learning algorithm for the high-level fuzzy Petri net (HLFPN) models in order to perform structure and parameter learning simultaneously. In addition to the HLFPN itself, the difference and similarity among a variety of subclasses concerning Petri nets are also discussed. As compared with the fuzzy adaptive learning control network (FALCON), the HLFPN model preserves the advantages that: 1) it offers more flexible learning capability because it is able to model both IF-THEN and IF-THEN-ELSE rules; 2) it allows multiple heterogeneous outputs to be drawn if they exist; 3) it offers a more compact data structure for fuzzy production rules so as to save information storage; and 4) it is able to learn faster due to its structural reduction. Finally, main results are presented in the form of seven propositions and supported by some experiments.

Correctness in hierarchical knowledge-based requirements.

As expert system technology gains broader acceptance, the need to build and maintain large-scale knowledge bases will assume greater importance. Traditional approaches to knowledge-based systems (KBSs) verification have generally adopted a pairwise comparison of rules, making them slow for large-scale KBSs. This paper introduces the least fixpoint semantics of a predicate/transition (pr/t) net model into the KBSs for the purposes of speeding up the computation and saving the design time of KBSs. An efficient fault diagnosis algorithm is presented to locate some fault(s) made in the KBS design. The significance of this work is that frame- and rule-based hardware description language (FARHDL) can easily form a KBS, and the pr/t net model provides a T-invariant technique to verify the correctness of KBS requirements. Thus, the performance of a computer-aided design (CAD) tool for digital systems can be improved to some extent.

Requirements specification and analysis of digital systems using fuzzy and marked Petri nets.

Fuzzy information often appears in the system requirements. Fuzzy Petri nets (FPN) are Petri nets in which certain fuzzy truth values are assigned to its transitions. We show how the FPN model can be used for formal specification and verification of digital systems. The consistent FPN model is actually a state machine, from which we can obtain a consistent marked Petri net (MPN) model. Based on the consistent MPN model, the hardware prototype at register transfer level can be easily induced by using the optimization rules. Finally, main results are presented in the form of three theorems and are supported by some experiments.