Message-Passing Algorithm for Sum-rate Maximization in Multi-antenna Underlay D2D Communications

Abstract

In this paper, we propose a resource allocation of device-to-device (D2D) communications underlaying multiantenna cellular systems by employing message-passing algorithm. It is possible to construct the factor graph of the D2D pairs by considering the interference power at the base station and among the D2D pairs. The goal is to avoid interfering the cellular transmission while increasing the sum data rate of D2D pairs. We show that the proposed resource allocation based on message passing algorithm allow us to almost reach the performance of the exhaustive search and outperforms the random allocation. Since the associated factor graph is not fully connected, the complexity of the algorithm is significantly reduced compared to the exhaustive search.

Energy Efficient Resource Allocation for Underlaying Multi-D2D Enabled Multiple-Antennas Communications

“…5, we illustrate cumulative density function (CDF) of GEE of the proposed joint algorithm by employing both MRC and MMSE receiver for underlaying D2D communications with K d = 8 and N r = 4. We compare the proposed solution with the one given in [23], that first applies Algorithm 1 and then uses MPA to maximize the sum data rate and with the optimal exhaustive search, where Algorithm 1 is followed by an exhaustive search on resource allocation for D2D aiming at maximizing the GEE. According to the performance results, the GEE of the proposed algorithm is significantly increased compared to the MPA for the sum data rate maximization and gives almost the same GEE results as the exhaustive search.…”

“…A joint channel and power allocation in a time-varying environment for underlaying D2D network has been addressed in [20] in a fully distributed approach. In order to maximize overall sum data rate for multi-D2D multiple antennas cellular systems, an interference management algorithm has been presented for the multiuser systems in [21] and Message Passing Algorithm (MPA) [22] based on the resource allocation in [23] has been examined. In contrast to these works, in this paper, we focus on the maximization of the GEE while employing multi-D2D pairs for each RB.…”