| With the development of technology and society, the network is used more andmore frequently. The life without network is unimaginable, so the development ofnetwork is increasingly affecting people’s life. Due to the expanding of size, thenetwork began to appear heterogeneous, intelligent and dynamic. The network ismore and more complex, which make it more difficult to diagnose the network fault.The paper summarizes the basic method of network fault diagnosis, and analyzesthe various type of network fault, based on which determine the basic steps ofnetwork. Meanwhile, the paper deeply analyzes fuzzy Petri net reasoning techniquesand its application in network fault diagnosis.In view of the existing problem of subjective random in the process of fuzzytoken determine, an effective method based on cloud model is proposed fordetermining fuzzy tokens combining cloud model and fuzzy Petri nets. Backwardcloud generator is employed to transform the expert or experimental knowledge intocloud concepts. And forward cloud generator is used to compute the degree ofcertainty that the evaluated object to the cloud concept. Thus the initial token of thefuzzy Petri nets is determined. The presented example confirms the feasibility andvalidity of cloud model in computing fuzzy tokens.By combing the simplified algorithm of fuzzy Petri net and a forward reasoningalgorithm, the paper presents an improved algorithm of fuzzy Petri net. Combiningthe fuzzy Petri nets and the diagnoses of network fault, the “Ethernet failure” is set asan example to explain the use of original fuzzy Petri net and the improved fuzzynetwork in the application of network fault diagnosis. The reasoning result of theimproved algorithm is the same with the backward reasoning algorithm, which showthat the improved reasoning algorithm can be effect used in the diagnosis of networkfault. Compared with the original fuzzy Petri network reasoning algorithm, it canreduce the reasoning scale and improve the reason speed. |
PDF Full Text Request