Long-term power allocation for multi-channel device-to-device communication based on limited feedback information

Ruhallah AliHemmati, Ben Liang, Min Dong, Gary Boudreau, S. Hossein Seyedmehdi

Abstract: In underlay Device-to-Device (D2D) communication, where a D2D pair reuses the cellular spectrum and creates interference to regular cellular users, there exists a tradeoff between the achieved D2D rate and the interference to cellular users. In this work, we present stochastic optimization solutions to allocate the D2D transmission power over multiple resource blocks (RBs), to maximize the D2D rate, under a sum-power constraint and long-term individual power constraints over each RB at the D2D transmitter, which gives probabilistic guarantees on the interference to regular cellular users. This stochastic optimization problem can be solved optimally in the Lagrange dual domain with stochastic subgradient updating. However, since the vector channel state space is exponentially increasing in size, the standard subgradient updating method has prohibitive computation and storage complexity. Instead, by first showing that only the signs of subgradients are necessary to find the optimal Lagrange multipliers, we propose a distributed algorithm where each interference-victim cellular user calculates the subgradient and reports only its sign in one-bit feedback. Simulation results demonstrate the effectiveness of the proposed optimization with limited feedback.