A Shifting Strategy for Locality Exploration in Channel Assignment

This work investigates minimizing the number of channels required for channel assignment in wireless cellular systems using cumulative cochannel interference constraints. The goal is to explore how easily solutions that are optimal for disjoint geographic regions can be combined into solutions that are simultaneously optimal for the entire collection of regions. The results are applicable to the design of distributed DCA heuristics. An Integer Programming (IP) model generates optimal solutions for individual regions. The IP formulation is enhanced so it can combine isomorphic equivalents of the local solutions into a global solution. This is accomplished by encoding channel shifts into linear constraints. The amount of information exchanged among regions is controlled by limiting the size of the neighborhood around a group for which it knows the number of channels used in other groups. The effectiveness of this strategy is evaluated for different neighborhood sizes, demand densities, and group formulation policies relevant in a distributed DCA context.



Results and Observations


K. Daniels, K. Chandra, Harish Rathi

DCA Home