A Benchmark for Evaluating Approximated Algorithms for Fair Wireless Channel Allocation

Abstract

We present a benchmark for the performance evaluation of heuristic and meta-heuristic approaches to fair distribution of indivisible goods. The specific problem reflected by the benchmark data sets is Wireless Channel Allocation (WCA), and the approach to fair distribution is to choose from feasible allocations by the maximum set of a fairness relation between their corresponding allocation performances. The effort for exhaustive search for such maximum sets is rapidly increasing and even problems with 10 users may already be beyond today's computing capabilities. Here we present the results for up to 7 users and also discuss some general aspects of using fairness relations in the prescribed manner, also indicating the efficiency of the approach in terms of establishing rather small maximum sets with much overlap of maximal elements among different fairness relations.