Hybrid Ant Colony Optimization for the Channel Assignment Problem in Wireless Communication

In this chapter, we investigate the channel assignment problem (CAP) that is critical in wireless communication applications. Researchers strive to develop algorithms that are able to effectively assign limited channels to a number of cells with nonhomogeneous demands. Inspired by the recent success of metaheuristics, a hybrid ant colony optimisation (HACO) is proposed in this chapter. The HACO embodies several problem-dependent heuristics including ordering, sequential packing, and a local optimiser into an ACO framework. The advantages of this hybrid are two-fold. First, the EMC constraints can be effectively handled by the problem-dependent heuristics instead of using a penalty function as observed in other works which may lengthen the elapsed time in order to reach convergence. Second, the embedded heuristics serve as intensification strategies conducted by the metaheuristic framework and help improve the generated solutions from different view points. The performance of the HACO algorithm is evaluated on the Philadelphia benchmark set, such that it can be compared to that of existing approaches. It is observed from the experimental results that the HACO algorithm can solve optimally six of the eight benchmark problems and obtain near-optimal solutions for the other two problems which have been known to be the most difficult in the literature. For practical reasons, we only allow the HACO algorithm to run for a relatively short time compared to that used by other approaches. It is plausible to get a better result if more computational time is allocated.