Matching-Coalition Based Cluster Formation for D2D Multicast Content Sharing

Yan Wu,Dan Wu,Lianxin Yang,Xin Shi,Liang Ao,Qinxue Fu

DOI: https://doi.org/10.1109/access.2019.2918942

IF: 3.9

2019-01-01

IEEE Access

Abstract:Device-to-device (D2D) multicast content sharing has become a promising technology to improve the services within a local area. This paper investigates the problem of cluster formation (i.e., cluster head (CH) selection and user grouping) for D2D multicast content sharing in cellular networks. Each cluster consists of one CH and multiple content requesters (CRs). First, the selected CHs should be able to serve as many CRs as possible. To achieve this, they should have higher physical-social centrality and should not stay physically close to each other. Second, the issue of user grouping is modeled as a social-aware rate maximization problem, where the social-aware rate is introduced to ensure the physical link quality and to stimulate an efficient cooperation between the CHs and CRs. In order to solve the optimization problem in a decentralized and low-complexity way, we propose a matching and coalition approach. Specifically, the optimization problem is transformed into a two-sided many-to-one matching game with externality, and a distributed user grouping algorithm is proposed to obtain a stable matching result. Considering the performance gap resulted from the externality, the problem is then transformed into a coalition formation game. Without loss of generality, a user grouping optimization algorithm with user transfer and merging is proposed. Moreover, some properties, e.g., the convergence, stability, and complexity, of the proposed algorithms are discussed. Finally, the performance of our proposed scheme is evaluated by the extensive simulations based on the comparisons with other schemes.