Device-to-device communications underlying UAV-Supported social networking

The integration of device-to-device (D2D) communications technology can further boost system performance for unmanned aerial vehicles supported pervasive social networking. In this paper, we maximize the sum social group utility of the social networking by power optimization, whereas the quality-of-service requirements of all D2D users and downlink users must be satisfied. In the proposed formulation, we consider the physical interference and social connections between users in the physical/social domain, respectively. However, the formulated problem is strongly NP-hard and intractable, to solve it we reveal the hidden non-convex components, and approximate the original problem with a sequence of more tractable convex problems. We propose an efficient semi-decentralized algorithm by decomposing the complex problem into several simple subproblems, which is proved to converge to the stationary solutions of the original problem. Simulation results show that our proposed method can achieve comparable performance but much less time complexity than the centralized method and obtain superior performance.