Extended knowledge attributed Petri Nets

Abstract

Choosing the best way for describing physical reality has always been standing in focus of research. Several methodologies have been developed based on classical mathematics, or statistics and also new disciplines — such as soft-computing techniques — appeared. Petri Nets as one of the most naturalistic modeling methodologies are well suited to describe complex process in general. However in some fields of modeling the describing power of basic Petri Nets proved not to be robust enough, therefore several extensions were made to the original concept. Colored tokens (Colored Petri Nets), stochastic delayed streaming of mobile entities (Stochastic Petri Nets), object oriented architecture (Object Oriented Petri Nets), numerical (Numerical Petri Nets) and linguistic attributes (Fuzzy Petri Nets) broaden the range of capabilities. In some fields of problem solving, usage of static and mobile knowledge bases is needed: e.g., flexible manufacturing systems, or intelligent traffic simulation. These problems to be investigated involved new conceptual developments of Petri Nets and led to the introduction of Knowledge Attributed Petri Nets. At the same time distributed control in simulation appeared, intelligent agents supported the connection of mobile knowledge bases and static inference engines in an effective way. The mentioned extensions brought general support in model synthesis, but some unsolved questions remained related to the implementation of intelligent mobile entities. This paper highlights a new level of AI controlled simulation introducing the Extended Knowledge Attributed Petri Nets that offer the capability of easy implementation of mobile inference engines and knowledge base, providing general mobile AI in Petri Nets.