Distributed algorithms for dynamic channel allocation in microcellular systems

The capacity of a microcellular system using dynamic channel allocation was studied. It was found that these systems can self-organize, with little loss in capacity, by using channel-allocation algorithms that are simple, practical, and local. The performances of both deterministic and probabilistic algorithms were calculated. Two classes of isotropic algorithms look particularly promising: the timid class, which is the simplest, and the aggressive class, which could provide improvements in system capacity and blocking probability. In particular, at the expense of additional rearrangements per call, these local algorithms can approach the capacity achieved by a global channel-allocation strategy.<>