Distributed WRBG Matching Approach for Multiflow Two-Way D2D Networks

Device-to-device (D2D) communication has great potential to improve spectrum efficiency and offload traffic for cellular networks. In this paper, we focus on a multiflow two-way D2D network with decode-and-forward (DF) relays, coexisting with OFDMA cellular network. The spectrum sharing and relay selection are considered to minimize the outage probability of the device in D2D networks. The induced problem is a complicated probabilistic integral programming. A novel weighted random bipartite graph (WRBG)-based minimum weight maximum matching (MWMM) approach will be proposed in this paper. To offload not only the traffic but also the signaling and computation overhead, the improved min-sum algorithm will be applied to find the MWMM in the distributed manner with only polynomial complexity. The proposed approach enjoys an advantage that the close-form approximation formulas for optimal outage probability and diversity-multiplexing tradeoff can be derived by analyzing the properties of MWMM on WRBG. Both the theoretical derivations and simulation results will illustrate that the proposed approach for multiflow two-way D2D networks achieves the same performance as single-flow two-way D2D systems. Therefore, the distributed WRBG matching approach yields not only a practical distributed algorithm, but also a simple and elegant theoretical framework for multiflow two-way D2D networks.