Optimal caching placement in D2D networks

Abstract

When we store a popular content everywhere in wireless caches, we do not always obtain an efficient gain in terms of overcoming the challenges of the limited backhaul. Furthermore limited size of devices'memories becomes also another challenge to study. As the last one is assumed often fixed, we think how and which files are placed in the caches. So, we need to find an optimal content placement. In this work, the main goal is to maximize user's hit probability. For this reason, we start by formulating an optimization problem that gives an optimal probabilistic content placement. This placement policy depends on time connection that characterizes user mobility in D2D network model. In order to study a probabilistic content placement (or caching placement), the locations of wireless terminals for our network model are superposed as an independent point process. In the beginning, we should formulate the expression of service success probability of a typical receiver requesting some particular content and which captures the system performance. In this paper, we consider mobiles users. Consequently, based on time connection assigned for each device, D2D links have two different states: either be activated or deactivated.