Decentralized inter-cell interference coordination by autonomous spectral reuse decisions

Future wireless packet switched cellular networks will require dense frequency reuse to achieve high capacity. At the same time, measures are required which limit the interference on the frequency carriers. It is assumed that central entities performing the task of interference coordination with global knowledge should be avoided. Rather, distributed algorithms are sought for. To this end, we propose decentralized resource allocation algorithms that enable base stations to select a pool of favorable resources with low interference based on local knowledge only. The actual user-level resource allocation from that pool will then be performed by fast schedulers operating on the preselected resources within each cell. We analyze and evaluate the proposed resource selection algorithms by introducing a simplified wireless network model and applying methods from game theory. Proving the existence of Nash equilibria shows that stable resource allocations can be reached by selfish agents. In addition to that, we perform simulations to determine the speed of convergence and the resulting equilibrium interference levels. By comparing these to an optimal global solution, which is derived by solving an integer linear program, we are able to quantify the efficiency loss of the distributed game approach. It turns out that even though the distributed game results are sub-optimal, the low degree of system complexity and the inherent adaptability make the decentralized approach promising especially for dynamic scenarios.