Algorithm portfolios for noisy optimization: Compare solvers early

Marie-Liesse CAUWET; Jialin LIU; Olivier TEYTAUD

doi:10.1007/978-3-319-09584-4_1

Algorithm portfolios for noisy optimization: Compare solvers early

Marie-Liesse CAUWET, Jialin LIU, Olivier TEYTAUD^*

^*Corresponding author for this work

Research output: Book Chapters | Papers in Conference Proceedings › Conference paper (refereed) › Referred Conference Paper › peer-review

8 Citations (Scopus)

Abstract

Noisy optimization is the optimization of objective functions corrupted by noise. A portfolio of algorithms is a set of algorithms equipped with an algorithm selection tool for distributing the computational power among them. We study portfolios of noisy optimization solvers, show that different settings lead to different performances, obtain mathematically proved performance (in the sense that the portfolio performs nearly as well as the best of its' algorithms) by an ad hoc selection algorithm dedicated to noisy optimization. A somehow surprising result is that it is better to compare solvers with some lag; i.e., recommend the current recommendation of the best solver, selected from a comparison based on their recommendations earlier in the run.

Original language	English
Title of host publication	Learning and Intelligent Optimization : 8th International Conference, Lion 8, Revised Selected Papers
Editors	Panos M. PARDALOS, Mauricio G.C. RESENDE, Chrysafis VOGIATZIS, Jose L. WALTEROS
Publisher	Springer-Verlag Italia Srl
Pages	1-15
Number of pages	15
ISBN (Electronic)	9783319095844
ISBN (Print)	9783319095837
DOIs	https://doi.org/10.1007/978-3-319-09584-4_1
Publication status	Published - 2014
Externally published	Yes
Event	8th International Conference on Learning and Intelligent OptimizatioN, LION 2014 - Gainesville, FL, United States Duration: 16 Feb 2014 → 21 Feb 2014

Publication series

Name	Lecture Notes in Computer Science
Publisher	Springer
Volume	8426
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Name	Theoretical Computer Science and General Issues
Publisher	Springer
ISSN (Print)	2512-2010
ISSN (Electronic)	2512-2029

Conference

Conference	8th International Conference on Learning and Intelligent OptimizatioN, LION 2014
Country/Territory	United States
City	Gainesville, FL
Period	16/02/14 → 21/02/14

Access to Document

10.1007/978-3-319-09584-4_1

Cite this

CAUWET, M.-L., LIU, J., & TEYTAUD, O. (2014). Algorithm portfolios for noisy optimization: Compare solvers early. In P. M. PARDALOS, M. G. C. RESENDE, C. VOGIATZIS, & J. L. WALTEROS (Eds.), Learning and Intelligent Optimization : 8th International Conference, Lion 8, Revised Selected Papers (pp. 1-15). (Lecture Notes in Computer Science; Vol. 8426), (Theoretical Computer Science and General Issues). Springer-Verlag Italia Srl. https://doi.org/10.1007/978-3-319-09584-4_1

CAUWET, Marie-Liesse ; LIU, Jialin ; TEYTAUD, Olivier. / Algorithm portfolios for noisy optimization: Compare solvers early. Learning and Intelligent Optimization : 8th International Conference, Lion 8, Revised Selected Papers. editor / Panos M. PARDALOS ; Mauricio G.C. RESENDE ; Chrysafis VOGIATZIS ; Jose L. WALTEROS. Springer-Verlag Italia Srl, 2014. pp. 1-15 (Lecture Notes in Computer Science). (Theoretical Computer Science and General Issues).

@inproceedings{8e089fb651b3424b8d43031ac5164f55,

title = "Algorithm portfolios for noisy optimization: Compare solvers early",

abstract = "Noisy optimization is the optimization of objective functions corrupted by noise. A portfolio of algorithms is a set of algorithms equipped with an algorithm selection tool for distributing the computational power among them. We study portfolios of noisy optimization solvers, show that different settings lead to different performances, obtain mathematically proved performance (in the sense that the portfolio performs nearly as well as the best of its' algorithms) by an ad hoc selection algorithm dedicated to noisy optimization. A somehow surprising result is that it is better to compare solvers with some lag; i.e., recommend the current recommendation of the best solver, selected from a comparison based on their recommendations earlier in the run.",

author = "Marie-Liesse CAUWET and Jialin LIU and Olivier TEYTAUD",

year = "2014",

doi = "10.1007/978-3-319-09584-4_1",

language = "English",

isbn = "9783319095837",

series = "Lecture Notes in Computer Science",

publisher = "Springer-Verlag Italia Srl",

pages = "1--15",

editor = "PARDALOS, {Panos M.} and RESENDE, {Mauricio G.C.} and Chrysafis VOGIATZIS and WALTEROS, {Jose L.}",

booktitle = "Learning and Intelligent Optimization : 8th International Conference, Lion 8, Revised Selected Papers",

address = "Italy",

note = "8th International Conference on Learning and Intelligent OptimizatioN, LION 2014 ; Conference date: 16-02-2014 Through 21-02-2014",

}

CAUWET, M-L, LIU, J & TEYTAUD, O 2014, Algorithm portfolios for noisy optimization: Compare solvers early. in PM PARDALOS, MGC RESENDE, C VOGIATZIS & JL WALTEROS (eds), Learning and Intelligent Optimization : 8th International Conference, Lion 8, Revised Selected Papers. Lecture Notes in Computer Science, vol. 8426, Theoretical Computer Science and General Issues, Springer-Verlag Italia Srl, pp. 1-15, 8th International Conference on Learning and Intelligent OptimizatioN, LION 2014, Gainesville, FL, United States, 16/02/14. https://doi.org/10.1007/978-3-319-09584-4_1

Algorithm portfolios for noisy optimization: Compare solvers early. / CAUWET, Marie-Liesse; LIU, Jialin; TEYTAUD, Olivier.
Learning and Intelligent Optimization : 8th International Conference, Lion 8, Revised Selected Papers. ed. / Panos M. PARDALOS; Mauricio G.C. RESENDE; Chrysafis VOGIATZIS; Jose L. WALTEROS. Springer-Verlag Italia Srl, 2014. p. 1-15 (Lecture Notes in Computer Science; Vol. 8426), (Theoretical Computer Science and General Issues).

Research output: Book Chapters | Papers in Conference Proceedings › Conference paper (refereed) › Referred Conference Paper › peer-review

TY - GEN

T1 - Algorithm portfolios for noisy optimization: Compare solvers early

AU - CAUWET, Marie-Liesse

AU - LIU, Jialin

AU - TEYTAUD, Olivier

PY - 2014

Y1 - 2014

N2 - Noisy optimization is the optimization of objective functions corrupted by noise. A portfolio of algorithms is a set of algorithms equipped with an algorithm selection tool for distributing the computational power among them. We study portfolios of noisy optimization solvers, show that different settings lead to different performances, obtain mathematically proved performance (in the sense that the portfolio performs nearly as well as the best of its' algorithms) by an ad hoc selection algorithm dedicated to noisy optimization. A somehow surprising result is that it is better to compare solvers with some lag; i.e., recommend the current recommendation of the best solver, selected from a comparison based on their recommendations earlier in the run.

AB - Noisy optimization is the optimization of objective functions corrupted by noise. A portfolio of algorithms is a set of algorithms equipped with an algorithm selection tool for distributing the computational power among them. We study portfolios of noisy optimization solvers, show that different settings lead to different performances, obtain mathematically proved performance (in the sense that the portfolio performs nearly as well as the best of its' algorithms) by an ad hoc selection algorithm dedicated to noisy optimization. A somehow surprising result is that it is better to compare solvers with some lag; i.e., recommend the current recommendation of the best solver, selected from a comparison based on their recommendations earlier in the run.

UR - http://www.scopus.com/inward/record.url?scp=84905845235&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-09584-4_1

DO - 10.1007/978-3-319-09584-4_1

M3 - Conference paper (refereed)

AN - SCOPUS:84905845235

SN - 9783319095837

T3 - Lecture Notes in Computer Science

SP - 1

EP - 15

BT - Learning and Intelligent Optimization : 8th International Conference, Lion 8, Revised Selected Papers

A2 - PARDALOS, Panos M.

A2 - RESENDE, Mauricio G.C.

A2 - VOGIATZIS, Chrysafis

A2 - WALTEROS, Jose L.

PB - Springer-Verlag Italia Srl

T2 - 8th International Conference on Learning and Intelligent OptimizatioN, LION 2014

Y2 - 16 February 2014 through 21 February 2014

ER -

CAUWET ML, LIU J, TEYTAUD O. Algorithm portfolios for noisy optimization: Compare solvers early. In PARDALOS PM, RESENDE MGC, VOGIATZIS C, WALTEROS JL, editors, Learning and Intelligent Optimization : 8th International Conference, Lion 8, Revised Selected Papers. Springer-Verlag Italia Srl. 2014. p. 1-15. (Lecture Notes in Computer Science). (Theoretical Computer Science and General Issues). Epub 2014 Jul 31. doi: 10.1007/978-3-319-09584-4_1

Algorithm portfolios for noisy optimization: Compare solvers early

Abstract

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this