A Comprehensive Confirmation-based Selfish Node Detection Algorithm for Socially Aware Networks

Data transmission in socially aware networks is usually accomplished through opportunistic peer-to-peer links in a storage-carry-forward way. This way demands nodes to actively participate in forwarding cooperation. However, several nodes display selfish behavior due to resource constraints and other factors, refusing to expend their own resources to aid in forwarding data, which will result in reduced network performance. Thus, a comprehensive confirmation-based selfish node detection algorithm(CCSDA) is proposed in this paper. First, the algorithm constructs a reputation evaluation model, which defines the node’s communication satisfaction and energy trust using the node’s historical communication behavior and energy consumption state and quantifies the node’s direct reputation value from several dimensions. Considering the existence of some edge nodes that are misjudged due to low cooperation opportunities for objective reasons, node centrality is introduced in the energy trust assessment to improve detection accuracy. Combining the direct reputation, indirect reputation based on neighbor sharing, and historical comprehensive reputation determines the final comprehensive reputation value to judge the nodes’ selfish attributes. Secondly, to reduce the additional energy consumption induced by the acknowledgment message mechanism in the reputation detection mechanism, a high reputation threshold is set. For high reputation nodes above this threshold, the node behavior will be monitored utilizing the interaction information-based detection mechanism to mitigate the energy consumption and thus optimize the network performance and monitor the confirmatory selfish behavior by combining communication satisfaction and energy trust. Finally, to reduce the performance impact of the misclassification of the normal nodes in the reputation detection mechanism, the nodes that are determined to be abnormal by the reputation detection mechanism are evaluated twice using the interaction information-based detection mechanism to improve the detection rate. The experimental outcomes demonstrate that the algorithm optimizes the detection performance and reduces the performance loss caused by selfish nodes compared with existing detection algorithms.