Dynamic Incentive Design in Content Dissemination Process Through D2D Communication

With the development in machine learning area, future networks are better at predicting user demands. By proactively disseminating content into target users' caches through device-to-device communication, wireless traffic can be greatly offloaded. However, practical users are selfish and incentive needs are to be provided to encourage them to help spread the content. In this letter, we consider the incentive design in content dissemination process with awareness of link failure. A dynamic incentive mechanism is proposed based on Markov decision process and the optimal incentives are determined for different network states. Simulation results show that our dynamic scheme achieves a higher network utility than the optimal fixed incentive scheme with an average gain of 7.23% under information asymmetry and 48.23% with partial information.