Social-Aware Resource Allocation for Content Dissemination Networks: An Evolutionary Game Approach

Recently, the increasing popularity of local area services, such as YouTube, Facebook, and Twitter, on which thousands of clients subscribe and download popular contents all the while, has drawn more and more attention on the study of content dissemination networks. Device-to-device (D2D) communication, which allows two devices to directly communicate with each other, has become an effective content dissemination method. With the goal of reducing the delay of content dissemination process with D2D communication, we propose an evolutionary game (EG)-based distributed resource allocation scheme. Moreover, we theoretically prove the existence and stability of game equilibrium and further propose a global search algorithm to achieve equilibrium. In the algorithm, all D2D links will select resource adaptively based on the predicted contact duration Aiming at improving the prediction accuracy of contact duration, we propose social trajectory similarity (STS) to represent the overlapping of history trajectory among all the mobile users by mining user behavior patterns. Numerical results show that our proposed STS increases nearly 20% correlation with history trajectory overlap compared with jaccard index. Furthermore, our EG-based scheme reduces the delay over 15% compared with the coalition game scheme, and over 30% compared with the random selection scheme. In addition, our proposed scheme also increases throughput efficiency with nearly 20% and achieves better fairness.