The implementation of device-to-device (D2D) un-\udderlaying or overlaying pre-existing cellular networks has re-\udceived much attention due to the potential of enhancing the total\udcell throughput, reducing power consumption and increasing the\udinstantaneous data rate. In this paper we propose a distributed\udpower allocation scheme for D2D OFDMA communications and,\udin particular, we consider the two operating modes amenable\udto a distributed implementation: dedicated and reuse modes.\udThe proposed schemes address the problem of maximizing the\udusers’ sum rate subject to power constraints, which is known\udto be nonconvex and, as such, extremely difficult to be solved\udexactly. We propose here a fresh approach to this well-known\udproblem, capitalizing on the fact that the power allocation\udproblem can be modeled as a potential game. Exploiting the\udpotential games property of converging under better response\uddynamics, we propose two fully distributed iterative algorithms,\udone for each operation mode considered, where each user updates\udsequentially and autonomously its power allocation. Numerical\udresults, computed for several different user scenarios, show\udthat the proposed methods, which converge to one of the local\udmaxima of the objective function, exhibit performance close to\udthe maximum achievable optimum and outperform other schemes\udpresented in the literature.
展开▼
