Decentralized Content Dissemination in Fog Radio Access Network Using Unsupervised Learning Empowered Rate-Splitting Framework

07/07/2020 ∙ by Md. Zoheb Hassan, et al. ∙ 0

Multi-hop device-to-device (D2D) communication-aided decentralized content dissemination is investigated for a fog radio access network (F-RAN). In the proposed framework, two content-sharing D2D links establish a device-cluster. In each device-cluster, the content-holder device-users (DUs) transmit to the content-requester DUs via a relay fog user-equipment (F-UE) over the same radio resource blocks (RRBs). Such RRBs are shared with uplink F-RAN as well. Rate-splitting and common message decoding are used at each device-cluster. A multi-objective resource optimization, for device-clustering, device power allocation, and scheduling of RRBs and relay F-UEs, is devised to simultaneously maximize the overall capacity of D2D links and minimize transmission power of the active devices. The formulated optimization problem is solved in two steps. First, by utilizing two-dimensional principal component analysis based unsupervised-learning technique, a low-complexity device-clustering method is proposed. Second, a Stackelberg resource scheduling game is exploited to obtain the devices' power allocations and scheduling of RRBs and relay F-UEs among the device-clusters. A decentralized content dissemination framework, referred as rate-splitting for multi-hop D2D (RSMD), is developed. The convergence of the proposed RSMD framework to a Stackelberg-equilibrium and Pareto-efficient outcome is justified. Through extensive simulations, efficiency of the proposed RSMD framework is demonstrated.



There are no comments yet.


page 1

page 2

page 3

page 4

This week in AI

Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.