Robust Submodular RRH Selection for Joint Multicast Downlink Transmission

In future wireless systems, network densification is expected to pose an increasing challenge to cellular communication. User-centric concepts such as virtual cells in cloud radio access networks can exploit multi-connectivity, in order to avoid interruptions caused by handovers. In many cases the combinatorial problems related to assigning remote radio heads (RRHs) to virtual cells can be solved or approximated with worst-case guarantees using greedy submodular maximization. However, for multi-user-centric virtual cells, which simultaneously serve multiple users, this is no longer the case. In this paper, we propose a computationally efficient submodular RRH selection algorithm for multi-user-centric virtual cells with cooperative multicast beamforming in the downlink. This method aims at activating a small subset of multi-antenna RRHs, which maximizes the achievable multicast capacity when the beamforming vectors are jointly optimized over all active RRHs. Furthermore, we provide an upper bound on the optimal multicast capacity. Simulations show that the proposed method can achieve multicast capacities close to this upper bound, which is maximized by exhaustive search.