Applying the Fuzzy Set Theory to Electing and Workload－Sharing Mechanism of the Cluster Manager in Ad Hoc Networks

應用模糊理論於隨意式網路中管理者的選派與工作負載分攤機制

The mobile ad-hoc networks is combined, using high-mobility moving equipment. The overall performance of the dynamically changeable mobile ad hoc network is heavily influenced by the cost of frequent reassembling. Thus, the mechanism of electing a manager of a cluster makes more important for maintaining a stable network. In addition, the manager of a cluster in a heavy flow mobile ad hoc network will often be in an overloading situation. It will likely shorten the lifetime of the manager, as well as cause the network to congest. Therefore, in this paper, a mechanism for cluster manager electing and workload sharing is proposed. The decision to elect the cluster manager and backup cluster manager is also considered. They are determined by the fuzzy inference rules, which are built from experts’ experience. Thus, both the electing and the overloading problems of a manager are improved.