![]() | Only 14 pages are availabe for public view |
Abstract Most of the classes of Petri nets focus on modeling and analyzing dynamic behavior of static systems. The dynamic behavior as well as the static structure of the modeled system is governed by Petri net through the firing of its transition. Since the state and properties of an intelligent system are variable frequently (knowledge is updated or modified frequently), modeling approaches should capable of adjusting the model parameters according to the system dynamics (changes of the system). This thesis focuses on introducing a new class of Petri nets, adaptive fuzzy higher order Petri nets (AFHOPN), for modeling intelligent systems. It takes into account the changes of the weights, is capable of dynamically adjusting the parameters and has the learning ability. The proposed model has systematic procedure for supporting the fuzzy reasoning via two algorithms: reasoning algorithm for computing fuzzy beliefs, which requires the algebraic forms of the new state equation of the proposed model, and the other for updating the weights. We also gave the stability analysis of the proposed model. We expand this proposed model to apply for two applications: intelligent E-learning system and intelligent Computer Numerical Control (CNC) machines. |