Constrained base station clustering for cooperative post-cellular relay networks

We consider the downlink of a "post-cellular" network in which base stations (BSs) are no longer evenly spread in the entire area of service but rather placed where they can easily be installed. Accordingly, the BSs are not necessarily close to the users anymore but have to serve distant mobiles with aggressive spatial multiplexing, possibly with the help of additional low-complexity relays. In such networks, it is important to allow for flexible and dynamic cooperation between different infrastructure nodes to achieve high data rates and coverage. To this end, we develop a framework that optimizes the performance by dynamic BS clustering with joint transmission for one- and two-hop communication under practical conditions such as backhaul rate limitations and power control. The network performance is maximized by an evolutionary algorithm that optimizes the BS clustering, relay routing, and power allocation for a beamforming scheme with closed-form solution. We show that a very high performance can be achieved in an efficient way with small overhead and that high data rates can be delivered to areas without wired backhaul access.