Incentive Mechanism for P2P Networks Based on Feature Weighting and Game Theory

Received: 16 July 2019 Accepted: 29 November 2019 There are mainly three types of problems in Peer-to-Peer (P2P) networks such as free-riding nodes, group deception, and node bias. To solve this, the authors proposed an incentive mechanism for the P2P networks based on feature weighting and game theory. The mechanism first used the five comprehensive coefficients of node degree, node clustering coefficient, local clustering coefficient, all clustering coefficients, and correlation coefficients to form a feature clustering matrix through linear fusion; then, in order to maximize the overall revenue, a sparse matrix of revenue was generated through feature classification, noise reduction, mapping and iteration, highlighting the status of fully cooperative nodes; afterwards, combining group evolution and constraint rule sets, the authors determined group node dynamic adjustment rules, response service rules, and message query and forwarding rules, to maximize service efficiency; finally, a multilayered P2P dynamic service system was constructed to promote the active evolution of nodes, and avoid the negative migration. The simulation experiment was also performed to verity this mechanism. The research findings provide an effective idea for stimulating node behaviors, reducing negative node migration, and excellent node selection.