Comparing genetic algorithms and greedy heuristics for adaptation problems

The framework of evolutionary algorithms has proven to be quite successful in designing hybrid search algorithms. The quality of solutions achieved by such hybrid search engines is usually much better than the quality of solutions provided by any of the participating individual heuristics, but the price paid is slower computational speed. For problems where computational speed is crucial it is interesting to investigate the trade-off between the quality of solution and the time to reach it. This paper investigates one such problem, namely the adaptation of a distributed file system.