A comparison of problem decomposition techniques for the FAP

Abstract

This paper proposes a problem decomposition approach to solve hard Frequency Assignment Problem instances with standard meta-heuristics. The proposed technique aims to divide the initial problem into a number of easier subproblems, solve them and then recompose the partial solutions into one of the original problem. We consider the COST-259 MI-FAP instances and other Cardiff University test problems in order to simulate larger and more realistic networks. For both benchmarks the standard implementations of meta-heuristics do not generally produce a satisfactory performance within reasonable times of execution. However, the decomposed assignment approach can improve their results, both in terms of solution quality and runtime.

Optimized network dimensioning and planning for WiMAX technology

“…This reference provides a good insight on the decomposition methods such as primal decomposition and dual decomposition. As well as explore general decomposition structures and associated decomposition methods.With application to wireless network planning,[43] proposes a problem decomposition approach to solve hard Frequency Assignment Problem instances with standard meta-heuristics. The results show that the use of the decomposed assignment approach proves to be a very effective technique to solve large practical data sets using meta-heuristics when time is limited.…”