Cooperative Transmission and Probabilistic Caching for Clustered D2D Networks

Abstract

In this paper, we aim at maximizing the cache off loading gain for a clustered device-to-device (D2D) caching network by exploiting probabilistic caching and cooperative transmission among the cluster devices. Devices with surplus memory probabilistically cache a content from a known library. A requested content is either brought from the device’s local cache, cooperatively transmitted from catering devices, or downloaded from the macro base station as a last resort. Using stochastic geometry, we derive a closed-form expression for the offloading gain and formulate the offloading maximization problem. In order to simplify the objective function and obtain analytically tractable expressions, we derive a lower bound on the offloading gain,for which a suboptimal solution is obtained when considering a special case. Results reveal that the obtained suboptimal solution can achieve up to12%increase in the offloading gain compared to the Zipf’s caching technique. Besides, we show that the spatial scaling parameters of the network, e.g., density of clusters and distance between devices in the same cluster, play a crucial role in identifying the tradeoff between the content diversity gain andthe cooperative transmission gain