Decentralized Precoding for Cache-Enabled Ultra-Dense Radio Access Networks

Abstract

This letter studies the fully decentralized optimization of the precoding matrices for the downlink of multiple-input-multiple-output cache-enabled ultra-dense radio access network. To avoid signalling exchange between coordinated edge radio access nodes (RANs), we first construct a virtual downlink cache-enabled ultra-dense radio access network. Following, a new concept of signal-to-interference-leakage-plus-noise ratio is defined. Based on this concept, a virtual weighted sum rate maximization problem is formulated to optimize the precoders. We then propose a computationally efficient decentralized optimization algorithm that achieves stationary point to the problem. Numerical results show that the proposed algorithm can achieve more than 96.15% of the performance of the centralized optimization without any signalling exchange between coordinated edge RANs for multiple-input single-output cases.