A weighted fuzzy petri net model for knowledge learning and reasoning

Fuzzy Petri net (FPN) theory can be used as a computational paradigm for intelligent systems. This specification and engineering language provides a graphical tool for visualization, communication and interpretation, and supports to manipulate imprecise and vague information. However, it is lack of adjustment (learning) mechanism being proposed to cope with potential numerical deficiencies of these models, and to adapt system inconstancy. This paper proposes a weighted approach which can overcome this shortage. A new FPN model with adaptive weights is proposed for knowledge learning and reasoning. Based on this model, fuzzy knowledge reasoning and weight learning algorithms are developed