New techniques for algorithm portfolio design

Matthew Streeter, Stephen F. Smith
Proceedings of the 24th Conference on Uncertainty in Artificial Intelligence, PMLR R6:519-527, 2008.

Abstract

We present and evaluate new techniques for designing algorithm portfolios. In our view, the problem has both a scheduling aspect and a machine learning aspect. Prior work has largely addressed one of the two aspects in isolation. Building on recent work on the scheduling aspect of the problem, we present a technique that addresses both aspects simultaneously and has attractive theoretical guarantees. Experimentally, we show that this technique can be used to improve the performance of state-of-the-art algorithms for Boolean satisfiability, zero-one integer programming, and A.I. planning.

Cite this Paper

BibTeX
@InProceedings{pmlr-vR6-streeter08a, title = {New techniques for algorithm portfolio design}, author = {Streeter, Matthew and Smith, Stephen F.}, booktitle = {Proceedings of the 24th Conference on Uncertainty in Artificial Intelligence}, pages = {519--527}, year = {2008}, editor = {McAllester, David A. and Myllymäki, Petri}, volume = {R6}, series = {Proceedings of Machine Learning Research}, month = {09--12 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/r6/main/assets/streeter08a/streeter08a.pdf}, url = {https://proceedings.mlr.press/r6/streeter08a.html}, abstract = {We present and evaluate new techniques for designing algorithm portfolios. In our view, the problem has both a scheduling aspect and a machine learning aspect. Prior work has largely addressed one of the two aspects in isolation. Building on recent work on the scheduling aspect of the problem, we present a technique that addresses both aspects simultaneously and has attractive theoretical guarantees. Experimentally, we show that this technique can be used to improve the performance of state-of-the-art algorithms for Boolean satisfiability, zero-one integer programming, and A.I. planning.}, note = {Reissued by PMLR on 09 October 2024.} }
Endnote
%0 Conference Paper %T New techniques for algorithm portfolio design %A Matthew Streeter %A Stephen F. Smith %B Proceedings of the 24th Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2008 %E David A. McAllester %E Petri Myllymäki %F pmlr-vR6-streeter08a %I PMLR %P 519--527 %U https://proceedings.mlr.press/r6/streeter08a.html %V R6 %X We present and evaluate new techniques for designing algorithm portfolios. In our view, the problem has both a scheduling aspect and a machine learning aspect. Prior work has largely addressed one of the two aspects in isolation. Building on recent work on the scheduling aspect of the problem, we present a technique that addresses both aspects simultaneously and has attractive theoretical guarantees. Experimentally, we show that this technique can be used to improve the performance of state-of-the-art algorithms for Boolean satisfiability, zero-one integer programming, and A.I. planning. %Z Reissued by PMLR on 09 October 2024.
APA
Streeter, M. & Smith, S.F.. (2008). New techniques for algorithm portfolio design. Proceedings of the 24th Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research R6:519-527 Available from https://proceedings.mlr.press/r6/streeter08a.html. Reissued by PMLR on 09 October 2024.

Related Material