Robust Multiarmed Bandit Problems

The multiarmed bandit problem is a popular framework for studying the exploration versus exploitation trade-off. Recent applications include dynamic assortment design, Internet advertising, dynamic pricing, and the control of queues. The standard mathematical formulation for a bandit problem makes t...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:Management science Ročník 62; číslo 1; s. 264 - 285
Hlavní autori: Kim, Michael Jong, Lim, Andrew E.B.
Médium: Journal Article
Jazyk:English
Vydavateľské údaje: Linthicum Institute for Operations Research and the Management Sciences 01.01.2016
Predmet:
ISSN:0025-1909, 1526-5501
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Abstract The multiarmed bandit problem is a popular framework for studying the exploration versus exploitation trade-off. Recent applications include dynamic assortment design, Internet advertising, dynamic pricing, and the control of queues. The standard mathematical formulation for a bandit problem makes the strong assumption that the decision maker has a full characterization of the joint distribution of the rewards, and that ″arms″ under this distribution are independent. These assumptions are not satisfied in many applications, and the outof-sample performance of policies that optimize a misspecified model can be poor. Motivated by these concerns, we formulate a robust bandit problem in which a decision maker accounts for distrust in the nominal model by solving a worst-case problem against an adversary (″nature″) who has the ability to alter the underlying reward distribution and does so to minimize the decision maker's expected total profit. Structural properties of the optimal worst-case policy are characterized by using the robust Bellman (dynamic programming) equation, and arms are shown to be no longer independent under nature's worst-case response. One implication of this is that index policies are not optimal for the robust problem, and we propose, as an alternative, a robust version of the Gittins index. Performance bounds for the robust Gittins index are derived by using structural properties of the value function together with ideas from stochastic dynamic programming duality. We also investigate the performance of the robust Gittins index policy when applied to a Bayesian webpage design problem. In the presence of model misspecification, numerical experiments show that the robust Gittins index policy not only outperforms the classical Gittins index policy, but also substantially reduces the variability in the out-of-sample performance.
AbstractList The multiarmed bandit problem is a popular framework for studying the exploration versus exploitation trade-off. Recent applications include dynamic assortment design, Internet advertising, dynamic pricing, and the control of queues. The standard mathematical formulation for a bandit problem makes the strong assumption that the decision maker has a full characterization of the joint distribution of the rewards, and that "arms" under this distribution are independent. These assumptions are not satisfied in many applications, and the out-of- sample performance of policies that optimize a misspecified model can be poor. Motivated by these concerns, we formulate a robust bandit problem in which a decision maker accounts for distrust in the nominal model by solving a worst-case problem against an adversary ("nature") who has the ability to alter the underlying reward distribution and does so to minimize the decision maker's expected total profit. Structural properties of the optimal worst-case policy are characterized by using the robust Bellman (dynamic programming) equation, and arms are shown to be no longer independent under nature's worst-case response. One implication of this is that index policies are not optimal for the robust problem, and we propose, as an alternative, a robust version of the Gittins index. Performance bounds for the robust Gittins index are derived by using structural properties of the value function together with ideas from stochastic dynamic programming duality. We also investigate the performance of the robust Gittins index policy when applied to a Bayesian webpage design problem. In the presence of model misspecification, numerical experiments show that the robust Gittins index policy not only outperforms the classical Gittins index policy, but also substantially reduces the variability in the out-of-sample performance.
The multiarmed bandit problem is a popular framework for studying the exploration versus exploitation trade-off. Recent applications include dynamic assortment design, Internet advertising, dynamic pricing, and the control of queues. The standard mathematical formulation for a bandit problem makes the strong assumption that the decision maker has a full characterization of the joint distribution of the rewards, and that ″arms″ under this distribution are independent. These assumptions are not satisfied in many applications, and the outof-sample performance of policies that optimize a misspecified model can be poor. Motivated by these concerns, we formulate a robust bandit problem in which a decision maker accounts for distrust in the nominal model by solving a worst-case problem against an adversary (″nature″) who has the ability to alter the underlying reward distribution and does so to minimize the decision maker's expected total profit. Structural properties of the optimal worst-case policy are characterized by using the robust Bellman (dynamic programming) equation, and arms are shown to be no longer independent under nature's worst-case response. One implication of this is that index policies are not optimal for the robust problem, and we propose, as an alternative, a robust version of the Gittins index. Performance bounds for the robust Gittins index are derived by using structural properties of the value function together with ideas from stochastic dynamic programming duality. We also investigate the performance of the robust Gittins index policy when applied to a Bayesian webpage design problem. In the presence of model misspecification, numerical experiments show that the robust Gittins index policy not only outperforms the classical Gittins index policy, but also substantially reduces the variability in the out-of-sample performance.
The multiarmed bandit problem is a popular framework for studying the exploration versus exploitation trade-off. Recent applications include dynamic assortment design, Internet advertising, dynamic pricing, and the control of queues. The standard mathematical formulation for a bandit problem makes the strong assumption that the decision maker has a full characterization of the joint distribution of the rewards, and that "arms" under this distribution are independent. These assumptions are not satisfied in many applications, and the out-of- sample performance of policies that optimize a misspecified model can be poor. Motivated by these concerns, we formulate a robust bandit problem in which a decision maker accounts for distrust in the nominal model by solving a worst-case problem against an adversary ("nature") who has the ability to alter the underlying reward distribution and does so to minimize the decision maker's expected total profit. Structural properties of the optimal worst-case policy are characterized by using the robust Bellman (dynamic programming) equation, and arms are shown to be no longer independent under nature's worst-case response. One implication of this is that index policies are not optimal for the robust problem, and we propose, as an alternative, a robust version of the Gittins index. Performance bounds for the robust Gittins index are derived by using structural properties of the value function together with ideas from stochastic dynamic programming duality. We also investigate the performance of the robust Gittins index policy when applied to a Bayesian webpage design problem. In the presence of model misspecification, numerical experiments show that the robust Gittins index policy not only outperforms the classical Gittins index policy, but also substantially reduces the variability in the out-of-sample performance. Keywords: bandit problems; robust control; model uncertainty; relative entropy; games against nature History: Received February 28, 2013; accepted November 30, 2014, by Dimitris Bertsimas, optimization. Published online in Articles in Advance August 5, 2015.
The multiarmed bandit problem is a popular framework for studying the exploration versus exploitation trade-off. Recent applications include dynamic assortment design, Internet advertising, dynamic pricing, and the control of queues. The standard mathematical formulation for a bandit problem makes the strong assumption that the decision maker has a full characterization of the joint distribution of the rewards, and that “arms” under this distribution are independent. These assumptions are not satisfied in many applications, and the out-of-sample performance of policies that optimize a misspecified model can be poor. Motivated by these concerns, we formulate a robust bandit problem in which a decision maker accounts for distrust in the nominal model by solving a worst-case problem against an adversary (“nature”) who has the ability to alter the underlying reward distribution and does so to minimize the decision maker’s expected total profit. Structural properties of the optimal worst-case policy are characterized by using the robust Bellman (dynamic programming) equation, and arms are shown to be no longer independent under nature’s worst-case response. One implication of this is that index policies are not optimal for the robust problem, and we propose, as an alternative, a robust version of the Gittins index. Performance bounds for the robust Gittins index are derived by using structural properties of the value function together with ideas from stochastic dynamic programming duality. We also investigate the performance of the robust Gittins index policy when applied to a Bayesian webpage design problem. In the presence of model misspecification, numerical experiments show that the robust Gittins index policy not only outperforms the classical Gittins index policy, but also substantially reduces the variability in the out-of-sample performance. This paper was accepted by Dimitris Bertsimas, optimization.
Audience Trade
Academic
Author Kim, Michael Jong
Lim, Andrew E.B.
Author_xml – sequence: 1
  givenname: Michael Jong
  surname: Kim
  fullname: Kim, Michael Jong
– sequence: 2
  givenname: Andrew E.B.
  surname: Lim
  fullname: Lim, Andrew E.B.
BookMark eNp1kctLAzEQh4MoWB9Xb0JF8OTWSTbZpEctvkBRRM8hm01qyj5qkgX9702pqJVKIIHh-zLD_HbQZtu1BqEDDCNMBD9r2qBHBDAbEczyDTTAjBQZY4A30QCAsAyPYbyNdkKYAQAXvBigo6eu7EMc3vd1dMo3phpeqLZycfjou7I2TdhDW1bVwex_vbvo5eryeXKT3T1c307O7zLNAWImCqA5sRobogTnmmFjqRECBGWqSvWKMG4tAdCCKSxwKSwvNRc6CaWBfBcdL_-d--6tNyHKWdf7NrWUpOACQ8Go-KGmqjbStbaLXunGBS3PKWV5wSnkicrWUFPTGq_qtDXrUnmFH63h06lM4_Ra4WRFSEw073Gq-hDkKnj6C0yrdq0J6Qpu-hrDkl83iPZdCN5YOfeuUf5DYpCLlOUiZblIWS5STgL9I2gXVXRpHq9c_b92uNRmIXb-uwnNRc4gBfkJCjOzBQ
CitedBy_id crossref_primary_10_1109_TAC_2017_2747405
crossref_primary_10_1080_24725854_2021_1973156
crossref_primary_10_1287_opre_2020_0464
crossref_primary_10_1109_TAC_2017_2731817
crossref_primary_10_1287_opre_2021_0393
crossref_primary_10_1287_moor_2019_1019
crossref_primary_10_1287_opre_2017_1631
crossref_primary_10_1016_j_ejor_2020_08_028
crossref_primary_10_1007_s00199_020_01328_3
crossref_primary_10_1287_opre_2021_2209
crossref_primary_10_1002_asmb_2355
crossref_primary_10_1111_poms_13239
crossref_primary_10_1287_opre_2019_1918
crossref_primary_10_1016_j_jmateco_2018_10_002
crossref_primary_10_1016_j_jmp_2024_102844
crossref_primary_10_1287_opre_2016_1495
crossref_primary_10_1016_j_ejor_2023_08_004
crossref_primary_10_1007_s10479_015_1965_7
crossref_primary_10_1016_j_ejor_2024_07_040
crossref_primary_10_1016_j_omega_2022_102780
crossref_primary_10_1287_opre_2022_2403
crossref_primary_10_1016_j_orl_2018_05_005
crossref_primary_10_3390_en17194883
crossref_primary_10_1017_jpr_2023_24
crossref_primary_10_1007_s10479_025_06821_3
crossref_primary_10_1007_s42488_020_00024_0
crossref_primary_10_1287_msom_2022_0325
crossref_primary_10_1287_opre_2020_2041
crossref_primary_10_1109_TAC_2019_2895253
Cites_doi 10.1109/TAC.2009.2031725
10.1109/TAC.1986.1104332
10.1287/opre.1050.0216
10.1016/0167-6911(90)90110-G
10.1016/S0167-6377(99)00016-4
10.1287/opre.1090.0796
10.1016/j.jet.2006.06.010
10.1109/9.85059
10.1137/050642885
10.1016/0196-8858(85)90002-8
10.1287/opre.1030.0065
10.1287/opre.1100.0866
10.1515/9781400829385
10.1016/j.jet.2005.06.006
10.1214/aos/1176344552
10.1287/moor.1040.0129
10.1287/educ.1063.0021
10.1111/j.2517-6161.1979.tb01068.x
10.1287/mnsc.1120.1518
10.1090/S0002-9904-1952-09620-8
10.1109/TAC.2015.2418672
10.1137/S0097539701398375
10.1109/9.847720
10.1007/s10898-012-9969-1
10.1111/j.2517-6161.1980.tb01111.x
10.1145/1273496.1273587
10.1109/SFCS.1995.492488
10.1137/S0363012992237273
10.1007/PL00011380
10.1016/j.orl.2012.08.010
10.1137/120878768
10.1007/BF01211853
10.1002/asmb.874
10.1109/TAC.1973.1100265
10.1287/moor.1080.0364
10.1002/9781118165904
10.1287/opre.2013.1164
10.1287/moor.23.4.769
10.1007/s10479-015-1965-7
10.1287/opre.1110.0999
10.2307/1426972
10.1016/j.ejor.2012.02.033
10.1016/S0022-0531(03)00097-8
10.1137/S0363012901383837
10.1287/mnsc.1060.0613
10.1111/j.1467-937X.2007.00464.x
10.1287/moor.21.2.257
10.1137/S0895479896298130
10.1287/moor.1120.0566
10.1017/CBO9780511546921
10.1111/j.2517-6161.1996.tb02080.x
10.1287/opre.1070.0385
10.1023/A:1013689704352
ContentType Journal Article
Copyright 2016 INFORMS
COPYRIGHT 2016 Institute for Operations Research and the Management Sciences
COPYRIGHT 2016 Institute for Operations Research and the Management Sciences
Copyright Institute for Operations Research and the Management Sciences Jan 2016
Copyright_xml – notice: 2016 INFORMS
– notice: COPYRIGHT 2016 Institute for Operations Research and the Management Sciences
– notice: COPYRIGHT 2016 Institute for Operations Research and the Management Sciences
– notice: Copyright Institute for Operations Research and the Management Sciences Jan 2016
DBID AAYXX
CITATION
N95
8BJ
FQK
JBE
DOI 10.1287/mnsc.2015.2153
DatabaseName CrossRef
Gale Business: Insights
International Bibliography of the Social Sciences (IBSS)
International Bibliography of the Social Sciences
International Bibliography of the Social Sciences
DatabaseTitle CrossRef
International Bibliography of the Social Sciences (IBSS)
DatabaseTitleList
International Bibliography of the Social Sciences (IBSS)



CrossRef

DeliveryMethod fulltext_linktorsrc
Discipline Business
EISSN 1526-5501
EndPage 285
ExternalDocumentID A445367403
10_1287_mnsc_2015_2153
43835004
GroupedDBID -~X
18M
29M
2AX
3EH
3R3
4.4
5GY
7WY
7X5
85S
8FL
8VB
AAAZS
AABCJ
AAIKC
AAMNW
AAWTO
AAXLS
ABAWQ
ABBHK
ABDNZ
ABIVO
ABKVW
ABLWH
ABPPZ
ABUWG
ABXSQ
ABYYQ
ABZEH
ACGFO
ACHJO
ACHQT
ACNCT
ACXJH
ADEPB
ADGDI
ADMHG
ADNFJ
ADNWM
ADULT
AEGXH
AEILP
AEMOZ
AENEX
AEUPB
AFAIT
AFTQD
AGKTX
AHAJD
AHQJS
AIAGR
AKVCP
ALIPV
ALMA_UNASSIGNED_HOLDINGS
APTMU
ASMEE
AZQEC
BAAKF
BENPR
BEZIV
BPHCQ
CBXGM
CCKSF
CS3
CYVLN
DU5
DWQXO
EBA
EBE
EBO
EBR
EBS
EBU
EJD
F5P
FRNLG
FYUFA
GENNL
GNUQQ
GROUPED_ABI_INFORM_ARCHIVE
GROUPED_ABI_INFORM_RESEARCH
IAO
IEA
IOF
IPC
IPO
IPSME
IPY
ISL
ISM
ITC
JAAYA
JAV
JBC
JBMMH
JBZCM
JENOY
JHFFW
JKQEH
JLEZI
JLXEF
JPL
JPPEU
JST
K1G
K60
K6~
M0C
M2M
MV1
N95
NEJ
NIEAY
OFU
P2P
QWB
REX
RPU
SA0
SJN
TH9
TN5
U5U
WH7
XSW
XZL
Y99
YNT
YZZ
ZL0
.-4
0R1
1OL
41~
8AO
8FI
8FJ
AADHG
AAYJJ
AAYXX
ABDPE
ABEFU
ABUFD
ACTDY
ACYGS
ADBBV
AFFDN
AFFHD
AFFNX
AFKRA
AQUVI
BVXVI
CCPQU
CITATION
GUPYA
HGD
HVGLF
H~9
LPU
M0T
P-O
PHGZM
PHGZT
PJZUB
PPXIY
PQBIZ
PQBZA
PQQKQ
PROAC
PSYQQ
RNS
UKHRP
UKR
VOH
YYP
ZCG
8BJ
FQK
JBE
ID FETCH-LOGICAL-c700t-860432fc1e2a877c51ef4e880845adfc1d257ff200c85a181b8f7bc78ce2abe03
ISICitedReferencesCount 33
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000372114300015&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN 0025-1909
IngestDate Mon Nov 10 06:32:15 EST 2025
Mon Oct 20 22:15:29 EDT 2025
Sat Nov 29 11:28:17 EST 2025
Mon Oct 20 16:43:07 EDT 2025
Thu Oct 16 16:18:54 EDT 2025
Sat Nov 29 09:04:36 EST 2025
Tue Nov 18 21:19:39 EST 2025
Sat Nov 29 04:10:06 EST 2025
Thu May 29 13:25:43 EDT 2025
IsPeerReviewed true
IsScholarly true
Issue 1
Language English
LinkModel OpenURL
MergedId FETCHMERGED-LOGICAL-c700t-860432fc1e2a877c51ef4e880845adfc1d257ff200c85a181b8f7bc78ce2abe03
Notes ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
PQID 2678106548
PQPubID 40737
PageCount 22
ParticipantIDs proquest_journals_2678106548
gale_infotracmisc_A445367403
gale_infotracgeneralonefile_A445367403
gale_infotracacademiconefile_A445367403
gale_incontextgauss__A445367403
gale_businessinsightsgauss_A445367403
crossref_primary_10_1287_mnsc_2015_2153
crossref_citationtrail_10_1287_mnsc_2015_2153
jstor_primary_43835004
PublicationCentury 2000
PublicationDate 20160101
2016-01-00
PublicationDateYYYYMMDD 2016-01-01
PublicationDate_xml – month: 1
  year: 2016
  text: 20160101
  day: 1
PublicationDecade 2010
PublicationPlace Linthicum
PublicationPlace_xml – name: Linthicum
PublicationTitle Management science
PublicationYear 2016
Publisher Institute for Operations Research and the Management Sciences
Publisher_xml – name: Institute for Operations Research and the Management Sciences
References B20
B21
B22
B23
B25
B26
B27
B28
B29
B30
B31
B33
B34
B35
B36
B37
B39
Bertsekas DP (B9) 1995
B1
B2
White JM (B52) 2013
B3
B4
B5
B6
B7
B8
Kleinberg R (B32) 2004
B40
B41
B42
B43
B44
B45
B46
B47
B48
Lim AEB (B38) 2011; 21
B50
B51
B10
B54
B11
B55
B12
B56
B13
B57
B14
B58
B15
B16
B17
B18
B19
Gittins JC (B24) 1979; 41
Whittle P (B53) 1980; 42
References_xml – ident: B39
  doi: 10.1109/TAC.2009.2031725
– ident: B51
  doi: 10.1109/TAC.1986.1104332
– volume-title: Dynamic Programming and Optimal Control Volume II
  year: 1995
  ident: B9
– start-page: 697
  volume-title: Advances in Neural Information Processing Systems 17
  year: 2004
  ident: B32
– ident: B40
  doi: 10.1287/opre.1050.0216
– ident: B55
  doi: 10.1016/0167-6911(90)90110-G
– ident: B7
  doi: 10.1016/S0167-6377(99)00016-4
– ident: B13
  doi: 10.1287/opre.1090.0796
– ident: B27
  doi: 10.1016/j.jet.2006.06.010
– ident: B56
  doi: 10.1109/9.85059
– ident: B45
  doi: 10.1137/050642885
– ident: B33
  doi: 10.1016/0196-8858(85)90002-8
– ident: B11
  doi: 10.1287/opre.1030.0065
– ident: B46
  doi: 10.1287/opre.1100.0866
– ident: B28
  doi: 10.1515/9781400829385
– ident: B26
  doi: 10.1016/j.jet.2005.06.006
– ident: B20
  doi: 10.1214/aos/1176344552
– ident: B30
  doi: 10.1287/moor.1040.0129
– ident: B36
  doi: 10.1287/educ.1063.0021
– volume: 41
  start-page: 148
  issue: 2
  year: 1979
  ident: B24
  publication-title: J. Royal Statist. Soc. Ser. B
  doi: 10.1111/j.2517-6161.1979.tb01068.x
– ident: B37
  doi: 10.1287/mnsc.1120.1518
– ident: B44
  doi: 10.1090/S0002-9904-1952-09620-8
– ident: B58
  doi: 10.1109/TAC.2015.2418672
– ident: B4
  doi: 10.1137/S0097539701398375
– ident: B43
  doi: 10.1109/9.847720
– ident: B34
  doi: 10.1007/s10898-012-9969-1
– volume: 42
  start-page: 143
  issue: 2
  year: 1980
  ident: B53
  publication-title: J. Royal Statist. Soc. Ser. B
  doi: 10.1111/j.2517-6161.1980.tb01111.x
– ident: B42
  doi: 10.1145/1273496.1273587
– ident: B3
  doi: 10.1109/SFCS.1995.492488
– ident: B1
  doi: 10.1137/S0363012992237273
– ident: B8
  doi: 10.1007/PL00011380
– ident: B29
  doi: 10.1016/j.orl.2012.08.010
– ident: B5
  doi: 10.1137/120878768
– ident: B18
  doi: 10.1007/BF01211853
– ident: B48
  doi: 10.1002/asmb.874
– ident: B31
  doi: 10.1109/TAC.1973.1100265
– ident: B17
  doi: 10.1287/moor.1080.0364
– ident: B19
  doi: 10.1002/9781118165904
– ident: B12
  doi: 10.1287/opre.2013.1164
– ident: B6
  doi: 10.1287/moor.23.4.769
– ident: B15
  doi: 10.1007/s10479-015-1965-7
– ident: B47
  doi: 10.1287/opre.1110.0999
– ident: B54
  doi: 10.2307/1426972
– ident: B41
  doi: 10.1016/j.ejor.2012.02.033
– ident: B22
  doi: 10.1016/S0022-0531(03)00097-8
– ident: B25
  doi: 10.1137/S0363012901383837
– ident: B14
  doi: 10.1287/mnsc.1060.0613
– ident: B23
  doi: 10.1111/j.1467-937X.2007.00464.x
– ident: B10
  doi: 10.1287/moor.21.2.257
– ident: B21
  doi: 10.1137/S0895479896298130
– volume-title: Bandit Algorithms for Website Optimization: Developing, Deploying, Debugging
  year: 2013
  ident: B52
– ident: B57
  doi: 10.1287/moor.1120.0566
– ident: B16
  doi: 10.1017/CBO9780511546921
– volume: 21
  start-page: 643
  issue: 4
  year: 2011
  ident: B38
  publication-title: Math. Finance
– ident: B50
  doi: 10.1111/j.2517-6161.1996.tb02080.x
– ident: B35
  doi: 10.1287/opre.1070.0385
– ident: B2
  doi: 10.1023/A:1013689704352
SSID ssj0007876
Score 2.3914523
Snippet The multiarmed bandit problem is a popular framework for studying the exploration versus exploitation trade-off. Recent applications include dynamic assortment...
SourceID proquest
gale
crossref
jstor
SourceType Aggregation Database
Enrichment Source
Index Database
Publisher
StartPage 264
SubjectTerms Advertisements
Algorithms
Bayesian analysis
Decision making
Decisions
Distribution
Dynamic programming
Entropy (Information theory)
Experiments
Exploitation
Game theory
Indexes
Internet
Mathematical research
Online advertising
Profitability
Property
Rewards
Stochastic models
Tradeoff analysis
Title Robust Multiarmed Bandit Problems
URI https://www.jstor.org/stable/43835004
https://www.proquest.com/docview/2678106548
Volume 62
WOSCitedRecordID wos000372114300015&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
journalDatabaseRights – providerCode: PRVPQU
  databaseName: ABI/INFORM Collection
  customDbUrl:
  eissn: 1526-5501
  dateEnd: 20201213
  omitProxy: false
  ssIdentifier: ssj0007876
  issn: 0025-1909
  databaseCode: 7WY
  dateStart: 19870101
  isFulltext: true
  titleUrlDefault: https://www.proquest.com/abicomplete
  providerName: ProQuest
– providerCode: PRVPQU
  databaseName: ABI/INFORM Global (OCUL)
  customDbUrl:
  eissn: 1526-5501
  dateEnd: 20201213
  omitProxy: false
  ssIdentifier: ssj0007876
  issn: 0025-1909
  databaseCode: M0C
  dateStart: 19870101
  isFulltext: true
  titleUrlDefault: https://search.proquest.com/abiglobal
  providerName: ProQuest
– providerCode: PRVPQU
  databaseName: ProQuest Central (subscription)
  customDbUrl:
  eissn: 1526-5501
  dateEnd: 20201213
  omitProxy: false
  ssIdentifier: ssj0007876
  issn: 0025-1909
  databaseCode: BENPR
  dateStart: 19870101
  isFulltext: true
  titleUrlDefault: https://www.proquest.com/central
  providerName: ProQuest
– providerCode: PRVPQU
  databaseName: Psychology Database
  customDbUrl:
  eissn: 1526-5501
  dateEnd: 20201213
  omitProxy: false
  ssIdentifier: ssj0007876
  issn: 0025-1909
  databaseCode: M2M
  dateStart: 19870101
  isFulltext: true
  titleUrlDefault: https://www.proquest.com/psychology
  providerName: ProQuest
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwtV1Lb9QwELZQixAXxKtioZRFAnpYpc3LsXNsq624dKmqIsopchx7hcRmV-ss6s9nJnZeXZDKgYuVdWbjx-eMx874G0I-MCEkj2TusYJrL1aB9niYR15EIy0LP86lKOpgE2w24zc36aULxmnqcAKsLPntbbr6r1BDHoCNR2f_Ae72oZAB1wA6pAA7pPcC_mqZb0xlPQXFeoEWJh5dqSYudozp26Od98vEzYXdR_lFz60ed9nnrfOOvWV9ISfTo9P-1kFwd-ugc0dAj8YvK7V23neN01_rxNmrjFM5_Z1EjIgLdkXaV6pJuDV4nIa0pOVbmjvEvY_zRWmQVzKgR2CKRN0c1XyXn9mInHcosk_imEYJi5H1dTdkNEWFx759b6dk0EpJE7sXa-rYO6HM42GJA-vEzdHWT3Vrrq4NkOun5IlbOYxPLOLPyANVPiePmoMLL8h7C_y4A35sgR83wL8kX8-n12efPRcAw5PM9yuPJ0iYqGWgQsEZkzRQOlagcXlMRQH5BShcrWH0S04F2Go51yyXjGOUt1z50R7ZKZelekXGoc5TEaeaJorikj6N4aka7N1CQ1G-HhGvaXgmHTs8Bin5meEqEToqw47KsKMy7KgROWzlV5YX5a-SH7EfMxdUFRKD205mLjbGZB10I_KulkNSkhK9nqzAQOLQSegl1E0Kd4IEWogkZgPJTwPJuaVw_5Pg_kAQdKsc3N6rwW-biMS-FJQN_K8ZDZl7900WguEX4GFt_vqebX5DHnfv5T7ZqdYb9ZY8lL-qH2Z9UI_hA7J7Op1dXsGvC_8M0_DiN_64rfk
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=Robust+multiarmed+bandit+problems&rft.jtitle=Management+science&rft.au=Kim%2C+Michael+Jong&rft.au=Lim%2C+Andrew+E.B&rft.date=2016-01-01&rft.pub=Institute+for+Operations+Research+and+the+Management+Sciences&rft.issn=0025-1909&rft.volume=62&rft.issue=1&rft.spage=264&rft_id=info:doi/10.1287%2Fmnsc.2015.2153&rft.externalDBID=N95&rft.externalDocID=A445367403
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=0025-1909&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=0025-1909&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=0025-1909&client=summon