Based On The D-S Theory Of Reasoning Petri Net Applied In The Network Fault Diagnosis

With the wider use of the Internet, computer technology and network communication technology develop more quickly than ever before. Network Fault and the deterioration of network performance will become increasingly serious impact on the interests of users. The network fault diagnosis is mainly based on personal experience, and professional and technical personnel must remember lots of boring instructions. It is not very convenient in the actual operation. Some one time fault diagnosis is a long way from completion, and also caused enormous losses. Therefore, the pressing market developed a convenient and fast method of fault diagnosis technology to support the completion of the entire work. Fault can be found as soon as possible and reduce losses.This paper is based on basic theories of Petri net, associated with D-S theory of reasoning. This paper gives one new Petri net model which is based on the D-S theory of reasoning, and it is applied to fault diagnosis. Overall, the paper made the following points:1. Petri and the D-S theory of reasoning combined. D-S theory of reasoning is one kind of fuzzy inference theory, which can be applied to target identification, data integration, troubleshooting and other fields. The Petri net is one kind of good graph analysis tool, which has strict mathematics foundation and several kinds of abstract levels general network theory. It is especially suitable to express the dynamic system structure behavior and concurrent event. This paper makes them combined, and forms the new DSPN net. The places and transitions have certain credibility. According to the places'and transitions'credibility, we can get the unknown places'credibility.2. Simplify the system. The places in DSPN are set as one sample space, which focus on every element as an evidence,every transition as a rule, in which the parameters in uncertain reasoning are in a reciprocal relationship. Each place has a minimum credibility. If the credibility is smaller than the minimum, then we set the place's resource zero. Thus, under the condition that we have known the original system's DSPN model and the control rule, the whole system's node will reduce. Derivation process will be more easily attainable.3. Realization based on the modeling of error detection of DSPN. On the basis of DSPN, we integrated the approaches of detection of net error, and applied the DSPN theory to the detection of net error. First, make pairs of parts of D-s theory, Petri net and net error detection one by one. For instance: error-transition-rule, error occurrence -trigger of transition-application of rule and etc.On the basis of the above work, this paper demonstrated an example of the system's operation. Approve of the results shows that the proposed use of DSPN net for network error detection is a success, of which the method also reveals a vision of the effective and efficient use.