Optimal User Association for Massive MIMO Empowered Ultra-Dense Wireless Networks

TL;DR

This paper formulates and solves an optimal user-association problem in ultra-dense Massive MIMO networks, demonstrating significant improvements in network performance and optimal deployment strategies.

Contribution

It introduces a new optimization model for user association in Massive MIMO ultra-dense networks, addressing computational challenges and providing solutions for practical deployment.

Findings

Optimal user association significantly improves rate levels.

The proposed model achieves near-optimal solutions with reasonable complexity.

Deployment strategies greatly influence network performance.

Abstract

Ultra network densification and Massive MIMO are considered major 5G enablers since they promise huge capacity gains by exploiting proximity, spectral and spatial reuse benefits. Both approaches rely on increasing the number of access elements per user, either through deploying more access nodes over an area or increasing the number of antenna elements per access node. At the network-level, optimal user-association for a densely and randomly deployed network of Massive MIMO empowered access nodes must account for both channel and load conditions. In this paper we formulate this complex problem, report its computationally intractability and reformulate it to a plausible form, amenable to acquire a global optimal solution with reasonable complexity. We apply the proposed optimization model to typical ultra-dense outdoor small-cell setups and demonstrate: (i) the significant impact of optimal user-association to the achieved rate levels compared to a baseline strategy, and (ii) the optimality of alternative network access element deployment strategies.