Resource Allocation and User Association in User-Centric Dense mmWave Cellular Networks

User-centric overlapped clustering, relying on base station (BS) cooperation, is a promising architecture for densely deployed BSs in millimeter-wave (mmWave) networks. In this architecture, a user can be served by a set of cooperating BSs which reduces the interference received from neighboring BSs. This paper studies the problem of maximizing the number of served users in a dense mmWave network while guaranteeing the quality of service (QoS) required by each UE, defined by a received signal quality. Since the formulated problem is NP-hard, two near-optimal solutions are proposed that perform clustering and resource allocation. The first is a heuristic algorithm that builds the clusters by greedily associating the user with as many BSs as needed. The second approach is a binary particle swarm optimization (PSO) algorithm adapted to our constrained problem. Simulations confirm that the proposed algorithms approach the optimal solution with substantially lower computational complexity.