Energy Consumption in Cache-Enabled D2D Networks Based on Matérn Cluster Process

Abstract

Device-to-device (D2D) communications have significant advantages such as short range and low energy consumption. Caching at mobile user devices by leveraging D2D communications relieves the heavy load on backhaul links and enhances users’ quality of service. Therefore, in a cache-enabled clustered D2D network, we analyze the probability of successful transmission and propose a jointly optimizing bandwidth partitioning and caching policy strategy to solve the average energy consumption minimization problem. To simultaneously have the clustered and bounded characteristics of D2D communication in real hotspots areas, we use Matérn cluster process (MCP) to model the cache-enabled clustered D2D network. For a given D2D receiver (DR), we derive the successful transmission probability of getting its requested content. Moreover, leveraging the analysis of successful transmission probability, a joint bandwidth partitioning and caching policy optimization problem is formulated to reduce average energy consumption. Considering the complexity of directly solving the optimization problem, two sub-problems are decoupled and solved iteratively. The optimized solutions of bandwidth partitioning and caching policy sub-problems are obtained by taking the derivative and adaptive mutation particle swarm optimization (AMPSO) algorithm, respectively. The simulations show the influence of the related parameters on system performance and illustrate that the strategy to jointly optimize bandwidth partitioning and caching policy can significantly reduce energy consumption.