A Distributed SON-Based User-Centric Backhaul Provisioning Scheme

摘要: 5G definition and standardization projects are well underway, governing characteristics major challenges have been identified. A critical network element impacting the potential performance of networks is backhaul, which expected to expand in length breadth cater exponential growth small cells while offering high throughput order gigabit per second less than 1 ms latency with resilience energy efficiency. Such may only be possible direct optical fiber connections that often not available country-wide cumbersome expensive deploy. On other hand, a prime characteristic diversity, describes radio access network, also types user applications devices. Thus, we propose novel, distributed, self-optimized, end-to-end user-cell-backhaul association scheme intelligently associates users candidate based on corresponding dynamic backhaul conditions abiding by users’ requirements. Radio broadcast multiple bias factors, each reflecting indicator (DPI) such as capacity, latency, resilience, consumption, so on. given would employ these factors derive user-centric cell ranking motivates it select conforms Reinforcement learning used at optimise for DPI way maximise system minimising gap between achievable required quality experience (QoE). Preliminary results show considerable improvement QoE cumulative when compared state-of-the-art user-cell schemes.