|
|
A Benchmark Set Extension and Comparative Study for the HyFlex Framework Host Publication: 2015 IEEE Congress on Evolutionary Computation (CEC) Authors: S. Adriaensen, G. Ochoa and A. Nowé Publisher: IEEE Publication Date: May. 2015 Number of Pages: 8 ISBN: 978-1-4799-7492-4
Abstract: In this work we conduct a comparative study ofseveral publicly available, state-of-the-art hyper-heuristics forHyFlex in order to assess their generality across domains. Tothis purpose we extend the HyFlex benchmark set with 3 newproblem domains: The 0ǃ Knap Sack, Quadratic Assignmentand Max-Cut Problem. To our knowledge, this is the first publicextension of the benchmark since the CHeSC 2011 competition.In addition, this is the first study testing the Fair-Share IteratedLocal Search (FS-ILS) method, designed in prior research, usinga semi-automated design approach, on new unseen problemdomains. We show that, of the methods compared, Adap-HH(CHeSC 2011 winner) clearly perfoms the most consistently,overall. In addition, we identify a weakness of, as well as away to further simplify the FS-ILS method. Finally, we foundthat, overall, the state-of-the-art methods compared, generalizedmuch better than a naive baseline.
|
|
