The Univariate Marginal Distribution Algorithm Copes Well with Deception and Epistasis

In their recent work, Lehre and Nguyen (2019) show that the univariate marginal distribution algorithm (UMDA) needs time exponential in the parent populations size to optimize the DeceptiveLeadingBlocks (DLB) problem. They conclude from this result that univariate EDAs have difficulties with decepti...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:Evolutionary computation Ročník 29; číslo 4; s. 543
Hlavní autori: Doerr, Benjamin, Krejca, Martin S
Médium: Journal Article
Jazyk:English
Vydavateľské údaje: United States 01.12.2021
Predmet:
ISSN:1530-9304, 1530-9304
On-line prístup:Zistit podrobnosti o prístupe
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Abstract In their recent work, Lehre and Nguyen (2019) show that the univariate marginal distribution algorithm (UMDA) needs time exponential in the parent populations size to optimize the DeceptiveLeadingBlocks (DLB) problem. They conclude from this result that univariate EDAs have difficulties with deception and epistasis. In this work, we show that this negative finding is caused by the choice of the parameters of the UMDA. When the population sizes are chosen large enough to prevent genetic drift, then the UMDA optimizes the DLB problem with high probability with at most λ(n2+2elnn) fitness evaluations. Since an offspring population size λ of order nlogn can prevent genetic drift, the UMDA can solve the DLB problem with O(n2logn) fitness evaluations. In contrast, for classic evolutionary algorithms no better runtime guarantee than O(n3) is known (which we prove to be tight for the (1+1) EA), so our result rather suggests that the UMDA can cope well with deception and epistatis. From a broader perspective, our result shows that the UMDA can cope better with local optima than many classic evolutionary algorithms; such a result was previously known only for the compact genetic algorithm. Together with the lower bound of Lehre and Nguyen, our result for the first time rigorously proves that running EDAs in the regime with genetic drift can lead to drastic performance losses.
AbstractList In their recent work, Lehre and Nguyen (2019) show that the univariate marginal distribution algorithm (UMDA) needs time exponential in the parent populations size to optimize the DeceptiveLeadingBlocks (DLB) problem. They conclude from this result that univariate EDAs have difficulties with deception and epistasis. In this work, we show that this negative finding is caused by the choice of the parameters of the UMDA. When the population sizes are chosen large enough to prevent genetic drift, then the UMDA optimizes the DLB problem with high probability with at most λ(n2+2elnn) fitness evaluations. Since an offspring population size λ of order nlogn can prevent genetic drift, the UMDA can solve the DLB problem with O(n2logn) fitness evaluations. In contrast, for classic evolutionary algorithms no better runtime guarantee than O(n3) is known (which we prove to be tight for the (1+1) EA), so our result rather suggests that the UMDA can cope well with deception and epistatis. From a broader perspective, our result shows that the UMDA can cope better with local optima than many classic evolutionary algorithms; such a result was previously known only for the compact genetic algorithm. Together with the lower bound of Lehre and Nguyen, our result for the first time rigorously proves that running EDAs in the regime with genetic drift can lead to drastic performance losses.In their recent work, Lehre and Nguyen (2019) show that the univariate marginal distribution algorithm (UMDA) needs time exponential in the parent populations size to optimize the DeceptiveLeadingBlocks (DLB) problem. They conclude from this result that univariate EDAs have difficulties with deception and epistasis. In this work, we show that this negative finding is caused by the choice of the parameters of the UMDA. When the population sizes are chosen large enough to prevent genetic drift, then the UMDA optimizes the DLB problem with high probability with at most λ(n2+2elnn) fitness evaluations. Since an offspring population size λ of order nlogn can prevent genetic drift, the UMDA can solve the DLB problem with O(n2logn) fitness evaluations. In contrast, for classic evolutionary algorithms no better runtime guarantee than O(n3) is known (which we prove to be tight for the (1+1) EA), so our result rather suggests that the UMDA can cope well with deception and epistatis. From a broader perspective, our result shows that the UMDA can cope better with local optima than many classic evolutionary algorithms; such a result was previously known only for the compact genetic algorithm. Together with the lower bound of Lehre and Nguyen, our result for the first time rigorously proves that running EDAs in the regime with genetic drift can lead to drastic performance losses.
In their recent work, Lehre and Nguyen (2019) show that the univariate marginal distribution algorithm (UMDA) needs time exponential in the parent populations size to optimize the DeceptiveLeadingBlocks (DLB) problem. They conclude from this result that univariate EDAs have difficulties with deception and epistasis. In this work, we show that this negative finding is caused by the choice of the parameters of the UMDA. When the population sizes are chosen large enough to prevent genetic drift, then the UMDA optimizes the DLB problem with high probability with at most λ(n2+2elnn) fitness evaluations. Since an offspring population size λ of order nlogn can prevent genetic drift, the UMDA can solve the DLB problem with O(n2logn) fitness evaluations. In contrast, for classic evolutionary algorithms no better runtime guarantee than O(n3) is known (which we prove to be tight for the (1+1) EA), so our result rather suggests that the UMDA can cope well with deception and epistatis. From a broader perspective, our result shows that the UMDA can cope better with local optima than many classic evolutionary algorithms; such a result was previously known only for the compact genetic algorithm. Together with the lower bound of Lehre and Nguyen, our result for the first time rigorously proves that running EDAs in the regime with genetic drift can lead to drastic performance losses.
Author Krejca, Martin S
Doerr, Benjamin
Author_xml – sequence: 1
  givenname: Benjamin
  surname: Doerr
  fullname: Doerr, Benjamin
  email: doerr@lix.polytechnique.fr
  organization: Laboratoire d'Informatique (LIX), CNRS, École Polytechnique, Institut Polytechnique de Paris, Palaiseau, France doerr@lix.polytechnique.fr
– sequence: 2
  givenname: Martin S
  surname: Krejca
  fullname: Krejca, Martin S
  email: martin.krejca@lip.fr
  organization: Sorbonne Université, CNRS, LIP6, Paris, Francemartin.krejca@lip.fr
BackLink https://www.ncbi.nlm.nih.gov/pubmed/34623434$$D View this record in MEDLINE/PubMed
BookMark eNpNkElPwzAUhC1URBe4cUY-cgnYz4kTH6u2LFIRlxaOkWO7rVFiBzsp4t8TsUicZjT65ulppmjkvDMIXVJyQymHW3NUvpQlISDYCZrQjJFEMJKO_vkxmsb4RghlQOgZGrOUA0tZOkEvm4PBW2ePMljZGfwkw946WeOljV2wVd9Z7_C83vtgu0ODF741Eb-ausYfQ4CXRpn2m5FO41U7tGS08Ryd7mQdzcWvztD2brVZPCTr5_vHxXydKCZolygowDBdkDSvMgAhgatdLlOiK53mJOeMFTudATWZ5hkI0JSrgmtOmdb5gM_Q9c_dNvj33sSubGxUw3fSGd_HErKCcMFyAQN69Yv2VWN02QbbyPBZ_m0BX_iMYl0
CitedBy_id crossref_primary_10_1016_j_tcs_2024_114622
crossref_primary_10_1007_s00453_022_00977_1
crossref_primary_10_1016_j_ic_2023_105125
crossref_primary_10_1016_j_tcs_2023_114074
crossref_primary_10_1002_admt_202402075
crossref_primary_10_1109_TEVC_2022_3229038
ContentType Journal Article
Copyright 2021 Massachusetts Institute of Technology.
Copyright_xml – notice: 2021 Massachusetts Institute of Technology.
DBID CGR
CUY
CVF
ECM
EIF
NPM
7X8
DOI 10.1162/evco_a_00293
DatabaseName Medline
MEDLINE
MEDLINE (Ovid)
MEDLINE
MEDLINE
PubMed
MEDLINE - Academic
DatabaseTitle MEDLINE
Medline Complete
MEDLINE with Full Text
PubMed
MEDLINE (Ovid)
MEDLINE - Academic
DatabaseTitleList MEDLINE - Academic
MEDLINE
Database_xml – sequence: 1
  dbid: NPM
  name: PubMed
  url: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed
  sourceTypes: Index Database
– sequence: 2
  dbid: 7X8
  name: MEDLINE - Academic
  url: https://search.proquest.com/medline
  sourceTypes: Aggregation Database
DeliveryMethod no_fulltext_linktorsrc
Discipline Engineering
Computer Science
EISSN 1530-9304
ExternalDocumentID 34623434
Genre Journal Article
GroupedDBID ---
.4S
.DC
0R~
36B
4.4
53G
5GY
5VS
6IK
AAJGR
AAKMM
AALFJ
AALMD
AAYFX
ABAZT
ABDBF
ABJNI
ACM
ACUHS
ADL
ADPZR
AEBYY
AENEX
AENSD
AFWIH
AFWXC
AIKLT
AKRVB
ALMA_UNASSIGNED_HOLDINGS
ARCSS
ASPBG
AVWKF
AZFZN
BDXCO
BEFXN
BFFAM
BGNUA
BKEBE
BPEOZ
CGR
CS3
CUY
CVF
DU5
EAP
EAS
EBC
EBD
EBS
ECM
ECS
EDO
EIF
EMB
EMK
EMOBN
EPL
EST
ESX
F5P
FEDTE
FNEHJ
GUFHI
HGAVV
HZ~
I-F
I07
IPLJI
JAVBF
LHSKQ
MCG
MINIK
NPM
O9-
OCL
P2P
PK0
RMI
SV3
TUS
W7O
ZWS
7X8
ABVLG
AEJOY
ID FETCH-LOGICAL-c391t-c282e3d8047b5229a26cf7a40dbd47076338fd521e5d65292d16c86d613dd79a2
IEDL.DBID 7X8
ISICitedReferencesCount 12
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000724318500005&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN 1530-9304
IngestDate Fri Jul 11 16:47:11 EDT 2025
Thu Apr 03 06:58:10 EDT 2025
IsDoiOpenAccess false
IsOpenAccess true
IsPeerReviewed true
IsScholarly true
Issue 4
Keywords Estimation-of-distribution algorithm
runtime analysis
univariate marginal distribution algorithm
epistasis
theory
Language English
License 2021 Massachusetts Institute of Technology.
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-c391t-c282e3d8047b5229a26cf7a40dbd47076338fd521e5d65292d16c86d613dd79a2
Notes ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 23
OpenAccessLink https://hal.science/hal-03277754
PMID 34623434
PQID 2580693792
PQPubID 23479
ParticipantIDs proquest_miscellaneous_2580693792
pubmed_primary_34623434
PublicationCentury 2000
PublicationDate 2021-12-01
PublicationDateYYYYMMDD 2021-12-01
PublicationDate_xml – month: 12
  year: 2021
  text: 2021-12-01
  day: 01
PublicationDecade 2020
PublicationPlace United States
PublicationPlace_xml – name: United States
PublicationTitle Evolutionary computation
PublicationTitleAlternate Evol Comput
PublicationYear 2021
SSID ssj0013201
Score 2.4212284
Snippet In their recent work, Lehre and Nguyen (2019) show that the univariate marginal distribution algorithm (UMDA) needs time exponential in the parent populations...
SourceID proquest
pubmed
SourceType Aggregation Database
Index Database
StartPage 543
SubjectTerms Algorithms
Biological Evolution
Epistasis, Genetic
Models, Genetic
Population Density
Title The Univariate Marginal Distribution Algorithm Copes Well with Deception and Epistasis
URI https://www.ncbi.nlm.nih.gov/pubmed/34623434
https://www.proquest.com/docview/2580693792
Volume 29
WOSCitedRecordID wos000724318500005&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText
inHoldings 1
isFullTextHit
isPrint
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwpV05T8MwFLaAMsBAoVzlkpFYo7qOY8cTqkorFqoOHN0ix3agUkkCKf39PKeu6IKExBJlyKWXd3zv0PsQusl4HFHNWJAqFQfMEhmkJNaQ80B4TblVXZ7VZBNiNIonEzn2BbfKj1WufGLtqE2hXY28Q6OYcIilkt6WH4FjjXLdVU-hsYkaIUAZZ5hist5FIH5fKgkk5O2rwXdOO3ahi0QlrikV_g4u6yAzbP738_bRnoeXuLfUhwO0YfMWaq6oG7C35BbaXdtDeIieQVmwG9GAxBmwJ3bst44tC9-5tbqeEQv3Zq_wxvnbO-4Xpa3wi53NsKvj4jvrh2Owyg0elA6TVtPqCD0NB4_9-8ATLgQahDcPNORfNjQxYSIFXCYV5ToTihGTGiYIuKIwzgwEfBsZHlFJTZfrmBuABMYIuPwYbeVFbk8RNpnlwhBppAhZpKiEU00MPNlmKZGkja5XckxAoV2XQuW2-KqSH0m20cnyZyTlcvNGEjJAayxkZ3-4-xztUDd_Uo-eXKBGBuZsL9G2Xsyn1edVrSlwHI0fvgFmQMky
linkProvider ProQuest
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=The+Univariate+Marginal+Distribution+Algorithm+Copes+Well+with+Deception+and+Epistasis&rft.jtitle=Evolutionary+computation&rft.au=Doerr%2C+Benjamin&rft.au=Krejca%2C+Martin+S&rft.date=2021-12-01&rft.issn=1530-9304&rft.eissn=1530-9304&rft.volume=29&rft.issue=4&rft.spage=543&rft_id=info:doi/10.1162%2Fevco_a_00293&rft.externalDBID=NO_FULL_TEXT
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1530-9304&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1530-9304&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1530-9304&client=summon