A fuzzy expert system for network fault management

The paper builds on recent work toward the simplification of fault management within communication networks using a fuzzy expert system. Based on the symptom description derived from the customers and network monitor, the developed system draws on the past experience of communication networks from trouble history record to indicate the degree to which special devices may be faulty on a prioritized basis (fault diagnosis). Moreover, by network managers' past experience, a few corresponding actions are proposed to release the fault (fault recovery). Since managers tend to use linguistic variables to describe the recovery actions, each action is assigned a degree of possibility. With the fuzzy set operation, the composition of relations between fault diagnosis and fault recovery is achieved. To demonstrate its effectiveness, the proposed system is applied to the fault management in our Laboratory. It was found that the fuzzy expert system can emulate network managers and provide substantial assistance in decision-making with respect to network fault management.