Cellular Channel Assignment: A New Localized and Distributed Strategy

As the use of mobile communications systems grows, the need arises for new and more efficient channel allocation techniques. The total number of available channels on a real-world network is in fact a scarce resource, and many assignment heuristics suffer from a clear lack of flexibility (this is the case of Fixed Channel Allocation), or from high computational and communication complexity (as with channel borrowing techniques). Performance can be improved by representing the system with an objective function whose minimum is associated with a good configuration; the various constraints appear as penalty terms in the function. The problem is thus reduced to the search for a minimum, that is often performed via heuristic algorithms like Hopfield neural networks, simulated annealing or reinforcement learning. These strategies usually require a central process to have global information and decide for all cells. We consider an objective-function formulation of the channel assignment problem that has been previously solved by search heuristics; we prove that the search time for the global minimum of the objective function is O(nlog n), and therefore there is no need for search techniques. Finally we show that the algorithm that arises from this formulation can be modified so that global knowledge and synchronization are no longer required, and we give its distributed version. By simulating a cellular network with mobile hosts on a hexagonal cell pattern with uniform call distribution, we show that our technique actually performs better than the best known algorithms.