| Recent years, people have conducted many researches on application of petri net on artificial intelligence. But, with the development of application, people found out that the traditional Petri net have a obvious shortcoming that it cannot describe inexact knowledge. However, there are a lots of problems are uncertain and difficult to presented by certain knowledge in the real world. So, people brought forward the definition of fuzzy petri net which can deal with uncertain knowledge while fuzzy mathematics appeared. Based on the basic definition of fuzzy Petri net (FPN), the knowledge representation of fuzzy production system is discussed. The relation between the representation of fuzzy production system and fuzzy Petri net is discussed and a transformation algorithm is presented. An improved model of fuzzy Petri net is proposed according to the specialty of representation of fuzzy knowledge.. In this model, the number of the nodes is reduced and the net system is simiplified in certain degree. The forward reasoning algorithm and the backward reasoning algorithm of the production system is brought forward. In the case of which the presented information is insufficient, the system can proceed with the default reasoning. A bi-directional reasoning algorithm based on the FPN model which combined the above two reasoning strategy is also presented. By analyzing the dynamic running mechanism of fuzzy Petri net, a consistency maintenance algorithm is brought forward. In the case of non-monotonic reasoning, the newly added knowledge is attached with more value while the old knowledge is also considered effective. With the object-oriented programming technology and the visual programming technique, a prototype system is implemented by using Visual C++ 6.0 under Win32 platform.
|
PDF Full Text Request