Performance Assessment of Recursive Probability Matching for Adaptive Operator Selection in Differential Evolution

Mudita Sharma; Manuel López-Ibáñez; Dimitar Lubomirov Kazakov

doi:10.1007/978-3-319-99259-4_26

Performance Assessment of Recursive Probability Matching for Adaptive Operator Selection in Differential Evolution

Mudita Sharma, Manuel López-Ibáñez, Dimitar Lubomirov Kazakov

Computer Science

Research output: Chapter in Book/Report/Conference proceeding › Chapter

Abstract

Probability Matching is one of the most successful methods for adaptive operator selection (AOS), that is, online parameter control, in evolutionary algorithms. In this paper, we propose a variant of Probability Matching, called Recursive Probability Matching (RecPM-AOS), that estimates reward based on progress in past generations and estimates quality based on expected quality of possible selection of operators in the past. We apply RecPM-AOS to the online selection of mutation strategies in differential evolution (DE) on the bbob benchmark functions. The new method is compared with two AOS methods, namely, PM-AdapSS, which utilises probability matching with relative fitness improvement, and F-AUC, which combines the concept of area under the curve with a multi-arm bandit algorithm. Experimental results show that the new tuned RecPM-AOS method is the most effective at identifying the best mutation strategy to be used by DE in solving most functions in bbob among the AOS methods.

Original language	English
Title of host publication	15th Intl Conf. on Parallel Problem Solving from Nature
Subtitle of host publication	(PPSN 2018)
Publisher	Springer
Pages	321-333
Number of pages	13
DOIs	https://doi.org/10.1007/978-3-319-99259-4_26
Publication status	Published - Sept 2018

Publication series

Name	LNCS

Bibliographical note

© Springer Nature Switzerland AG 2018. This is an author-produced version of the published paper. Uploaded in accordance with the publisher’s self-archiving policy. Further copying may not be permitted; contact the publisher for details.

Access to Document

10.1007/978-3-319-99259-4_26

paper_66 (1)
© Springer Nature Switzerland AG 2018. This is an author-produced version of the published paper. Uploaded in accordance with the publisher’s self-archiving policy. Further copying may not be permitted; contact the publisher for details.
Accepted author manuscript, 845 KBLicence: Unspecified

https://link.springer.com/book/10.1007%2F978-3-319-99259-4Licence: Unspecified

Dimitar Lubomirov Kazakov
- dimitar.kazakovyork.acuk
- Computer Science - Reader
Person: Academic

Cite this

@inbook{1b0708cce9864f07a84e38a6148ff0c1,

title = "Performance Assessment of Recursive Probability Matching for Adaptive Operator Selection in Differential Evolution",

abstract = "Probability Matching is one of the most successful methods for adaptive operator selection (AOS), that is, online parameter control, in evolutionary algorithms. In this paper, we propose a variant of Probability Matching, called Recursive Probability Matching (RecPM-AOS), that estimates reward based on progress in past generations and estimates quality based on expected quality of possible selection of operators in the past. We apply RecPM-AOS to the online selection of mutation strategies in differential evolution (DE) on the bbob benchmark functions. The new method is compared with two AOS methods, namely, PM-AdapSS, which utilises probability matching with relative fitness improvement, and F-AUC, which combines the concept of area under the curve with a multi-arm bandit algorithm. Experimental results show that the new tuned RecPM-AOS method is the most effective at identifying the best mutation strategy to be used by DE in solving most functions in bbob among the AOS methods.",

author = "Mudita Sharma and Manuel L{\'o}pez-Ib{\'a}{\~n}ez and Kazakov, {Dimitar Lubomirov}",

note = "{\textcopyright} Springer Nature Switzerland AG 2018. This is an author-produced version of the published paper. Uploaded in accordance with the publisher{\textquoteright}s self-archiving policy. Further copying may not be permitted; contact the publisher for details. ",

year = "2018",

month = sep,

doi = "10.1007/978-3-319-99259-4_26",

language = "English",

series = "LNCS",

publisher = "Springer",

pages = "321--333",

booktitle = "15th Intl Conf. on Parallel Problem Solving from Nature",

address = "Germany",

}

TY - CHAP

T1 - Performance Assessment of Recursive Probability Matching for Adaptive Operator Selection in Differential Evolution

AU - Sharma, Mudita

AU - López-Ibáñez, Manuel

AU - Kazakov, Dimitar Lubomirov

N1 - © Springer Nature Switzerland AG 2018. This is an author-produced version of the published paper. Uploaded in accordance with the publisher’s self-archiving policy. Further copying may not be permitted; contact the publisher for details.

PY - 2018/9

Y1 - 2018/9

N2 - Probability Matching is one of the most successful methods for adaptive operator selection (AOS), that is, online parameter control, in evolutionary algorithms. In this paper, we propose a variant of Probability Matching, called Recursive Probability Matching (RecPM-AOS), that estimates reward based on progress in past generations and estimates quality based on expected quality of possible selection of operators in the past. We apply RecPM-AOS to the online selection of mutation strategies in differential evolution (DE) on the bbob benchmark functions. The new method is compared with two AOS methods, namely, PM-AdapSS, which utilises probability matching with relative fitness improvement, and F-AUC, which combines the concept of area under the curve with a multi-arm bandit algorithm. Experimental results show that the new tuned RecPM-AOS method is the most effective at identifying the best mutation strategy to be used by DE in solving most functions in bbob among the AOS methods.

AB - Probability Matching is one of the most successful methods for adaptive operator selection (AOS), that is, online parameter control, in evolutionary algorithms. In this paper, we propose a variant of Probability Matching, called Recursive Probability Matching (RecPM-AOS), that estimates reward based on progress in past generations and estimates quality based on expected quality of possible selection of operators in the past. We apply RecPM-AOS to the online selection of mutation strategies in differential evolution (DE) on the bbob benchmark functions. The new method is compared with two AOS methods, namely, PM-AdapSS, which utilises probability matching with relative fitness improvement, and F-AUC, which combines the concept of area under the curve with a multi-arm bandit algorithm. Experimental results show that the new tuned RecPM-AOS method is the most effective at identifying the best mutation strategy to be used by DE in solving most functions in bbob among the AOS methods.

U2 - 10.1007/978-3-319-99259-4_26

DO - 10.1007/978-3-319-99259-4_26

M3 - Chapter

T3 - LNCS

SP - 321

EP - 333

BT - 15th Intl Conf. on Parallel Problem Solving from Nature

PB - Springer

ER -

Performance Assessment of Recursive Probability Matching for Adaptive Operator Selection in Differential Evolution

Abstract

Publication series

Bibliographical note

Access to Document

Profiles

Dimitar Lubomirov Kazakov

Cite this