Evolutionary Algorithms for 5G Multi-Tier Radio Access Network Planning

Abstract

With the ever-increasing traffic demand of wireless users, resulting from the huge deployment of Internet-of-Things (IoT) devices and the emergence of smart city applications requiring ultra-low latency networks, the Fifth Generation (5G) of cellular networks have been introduced as a revolutionary broadband technology to boost the quality of service of mobile users. In this paper, we investigate the planning process for a 5G radio access network having mmWave Micro Remote Radio Units (mRRUs) on top of sub-6 GHz Macro Remote Radio Units (MRRUs). We rely on proper channel models and link budgets as well as Urban Macro-cells (UMa) and Urban Micro-cells (UMi) characteristics to carefully formulate a 5G network planning optimization problem. We aim to jointly determine the minimum number of MRRUs and mRRUs to install and find their locations in a given geographical area while fulfilling coverage and user traffic demand constraints. In order to solve this planning process, we propose a two-step process where we first employ a low complexity meta-heuristic algorithm to optimize the locations of RRUs followed by an iterative elimination method to remove redundant cells. To evaluate the performances of this proposed approach, we conduct a comparative study using Accelerated Particle Swarm Optimization and Simulated Annealing. Simulations results using sub-6 GHz UMa and 28 GHz UMi demonstrate the ability of the proposed planning approach to achieve more than 98% coverage with minimum cell capacity outage rate, not exceeding the 2%, for different scenarios and illustrate the efficiency of the evolutionary algorithms in solving this NP-hard problem in reasonable running time.

Keywords

• 5G networks
• radio heterogeneous network planning
• network dimensioning
• swarm intelligence
• evolutionary algorithm