Depth-expurgation based dynamic trust evaluation algorithm for ad hoc networks

Due to the wireless nature of ad hoc networks, two nodes out of mutual communication range need intermediate nodes cooperative to implement applications, such as resources discovery and packets forwarding. Trust evaluation between two peers directly affects the security and performance of networks. In this paper, a novel trust evaluation model is proposed for ad hoc networks. In this model, based on the theories of directed acyclic graph and the notion of social control, several rules are introduced to weight recommendation paths. A trust evaluation algorithm (DEA) is also proposed. During trust evaluation process, the depth of directed graph for an ad hoc network decreases step by step. Simulations show that the proposed algorithm can improve the accuracy of evaluation and effectively mitigate malicious nodes' denigration attacks.