Blind Dynamic Resource Allocation in Closed Networks via Mirror Backpressure
We study the problem of maximizing payoff generated over a period of time in a general class of closed queueing networks with a finite, fixed number of supply units that circulate in the system. Demand arrives stochastically, and serving a demand unit (customer) causes a supply unit to relocate from...
Gespeichert in:
Veröffentlicht in: | Management science 2024-08, Vol.70 (8), p.5445-5462 |
---|---|
1. Verfasser: | |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
container_end_page | 5462 |
---|---|
container_issue | 8 |
container_start_page | 5445 |
container_title | Management science |
container_volume | 70 |
creator | Kanoria, Yash |
description | We study the problem of maximizing payoff generated over a period of time in a general class of closed queueing networks with a finite, fixed number of supply units that circulate in the system. Demand arrives stochastically, and serving a demand unit (customer) causes a supply unit to relocate from the “origin” to the “destination” of the customer. The key challenge is to manage the distribution of supply in the network. We consider general controls including customer entry control, pricing, and assignment. Motivating applications include shared transportation platforms and scrip systems. Inspired by the mirror descent algorithm for optimization and the backpressure policy for network control, we introduce a rich family of
mirror backpressure
(MBP) control policies. The MBP policies are simple and practical and crucially do not need any statistical knowledge of the demand (customer) arrival rates (these rates are permitted to vary in time). Under mild conditions, we propose MBP policies that are provably near optimal. Specifically, our policies lose at most
O
(
K
T
+
1
K
+
η
K
)
payoff per customer relative to the optimal policy that knows the demand arrival rates, where
K
is the number of supply units,
T
is the total number of customers over the time horizon, and
η
is the demand process’ average rate of change per customer arrival. An adaptation of MBP is found to perform well in numerical experiments based on data from NYC Cab.
This paper was accepted by Gabriel Weintraub, revenue management and market analytics.
Funding:
Y. Kanoria was supported by the National Science Foundation’s Division of Civil, Mechanical, and Manufacturing Innovation [Grant CMMI-1653477].
Supplemental Material:
The data files and online appendices are available at
https://doi.org/10.1287/mnsc.2023.4934
. |
doi_str_mv | 10.1287/mnsc.2023.4934 |
format | Article |
fullrecord | <record><control><sourceid>proquest_cross</sourceid><recordid>TN_cdi_proquest_journals_3095235588</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>3095235588</sourcerecordid><originalsourceid>FETCH-LOGICAL-c247t-18a4124b5cd75917acd99d0f3ba993357c9122ecc6fb4623d75ba55dabfdf7303</originalsourceid><addsrcrecordid>eNqFkM1LwzAYxoMoOKdXzwHPrflsmuM2P2EqiIK3kKYpZGuTmXTK_ntbKnj09F5-z_O8_AC4xCjHpBTXnU8mJ4jQnEnKjsAMc1JknCN8DGYIEZ5hieQpOEtpgxASpShmYL1sna_hzcHrzhn4alPYR2Phom2D0b0LHjoPV21ItobPtv8OcZvgl9PwycUYIlxqs91Fm9I-2nNw0ug22YvfOwfvd7dvq4ds_XL_uFqsM0OY6DNcaoYJq7ipBZdYaFNLWaOGVlpKSrkwEhNijSmaihWEDlSlOa911dSNoIjOwdXUu4vhc29TrzbD136YVBRJTijnZTlQ-USZGFKKtlG76DodDwojNRpTozE1GlOjsSEAp4A1wbv0h5dSElQw9DEg2YQ434TYpf8qfwBFT3h5</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>3095235588</pqid></control><display><type>article</type><title>Blind Dynamic Resource Allocation in Closed Networks via Mirror Backpressure</title><source>Informs</source><creator>Kanoria, Yash</creator><creatorcontrib>Kanoria, Yash</creatorcontrib><description>We study the problem of maximizing payoff generated over a period of time in a general class of closed queueing networks with a finite, fixed number of supply units that circulate in the system. Demand arrives stochastically, and serving a demand unit (customer) causes a supply unit to relocate from the “origin” to the “destination” of the customer. The key challenge is to manage the distribution of supply in the network. We consider general controls including customer entry control, pricing, and assignment. Motivating applications include shared transportation platforms and scrip systems. Inspired by the mirror descent algorithm for optimization and the backpressure policy for network control, we introduce a rich family of
mirror backpressure
(MBP) control policies. The MBP policies are simple and practical and crucially do not need any statistical knowledge of the demand (customer) arrival rates (these rates are permitted to vary in time). Under mild conditions, we propose MBP policies that are provably near optimal. Specifically, our policies lose at most
O
(
K
T
+
1
K
+
η
K
)
payoff per customer relative to the optimal policy that knows the demand arrival rates, where
K
is the number of supply units,
T
is the total number of customers over the time horizon, and
η
is the demand process’ average rate of change per customer arrival. An adaptation of MBP is found to perform well in numerical experiments based on data from NYC Cab.
This paper was accepted by Gabriel Weintraub, revenue management and market analytics.
Funding:
Y. Kanoria was supported by the National Science Foundation’s Division of Civil, Mechanical, and Manufacturing Innovation [Grant CMMI-1653477].
Supplemental Material:
The data files and online appendices are available at
https://doi.org/10.1287/mnsc.2023.4934
.</description><identifier>ISSN: 0025-1909</identifier><identifier>EISSN: 1526-5501</identifier><identifier>DOI: 10.1287/mnsc.2023.4934</identifier><language>eng</language><publisher>Linthicum: INFORMS</publisher><subject>Averages ; backpressure ; control of queueing networks ; Customer services ; Customers ; Management science ; mirror descent ; no-underflow constraint ; Operations management ; Operations research ; Optimization ; Queueing ; Queuing ; Resource allocation ; Stochastic models ; Supply & demand</subject><ispartof>Management science, 2024-08, Vol.70 (8), p.5445-5462</ispartof><rights>Copyright Institute for Operations Research and the Management Sciences Aug 2024</rights><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed><cites>FETCH-LOGICAL-c247t-18a4124b5cd75917acd99d0f3ba993357c9122ecc6fb4623d75ba55dabfdf7303</cites><orcidid>0000-0002-1759-1009 ; 0000-0002-7221-357X</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://pubsonline.informs.org/doi/full/10.1287/mnsc.2023.4934$$EHTML$$P50$$Ginforms$$H</linktohtml><link.rule.ids>314,780,784,3692,27924,27925,62616</link.rule.ids></links><search><creatorcontrib>Kanoria, Yash</creatorcontrib><title>Blind Dynamic Resource Allocation in Closed Networks via Mirror Backpressure</title><title>Management science</title><description>We study the problem of maximizing payoff generated over a period of time in a general class of closed queueing networks with a finite, fixed number of supply units that circulate in the system. Demand arrives stochastically, and serving a demand unit (customer) causes a supply unit to relocate from the “origin” to the “destination” of the customer. The key challenge is to manage the distribution of supply in the network. We consider general controls including customer entry control, pricing, and assignment. Motivating applications include shared transportation platforms and scrip systems. Inspired by the mirror descent algorithm for optimization and the backpressure policy for network control, we introduce a rich family of
mirror backpressure
(MBP) control policies. The MBP policies are simple and practical and crucially do not need any statistical knowledge of the demand (customer) arrival rates (these rates are permitted to vary in time). Under mild conditions, we propose MBP policies that are provably near optimal. Specifically, our policies lose at most
O
(
K
T
+
1
K
+
η
K
)
payoff per customer relative to the optimal policy that knows the demand arrival rates, where
K
is the number of supply units,
T
is the total number of customers over the time horizon, and
η
is the demand process’ average rate of change per customer arrival. An adaptation of MBP is found to perform well in numerical experiments based on data from NYC Cab.
This paper was accepted by Gabriel Weintraub, revenue management and market analytics.
Funding:
Y. Kanoria was supported by the National Science Foundation’s Division of Civil, Mechanical, and Manufacturing Innovation [Grant CMMI-1653477].
Supplemental Material:
The data files and online appendices are available at
https://doi.org/10.1287/mnsc.2023.4934
.</description><subject>Averages</subject><subject>backpressure</subject><subject>control of queueing networks</subject><subject>Customer services</subject><subject>Customers</subject><subject>Management science</subject><subject>mirror descent</subject><subject>no-underflow constraint</subject><subject>Operations management</subject><subject>Operations research</subject><subject>Optimization</subject><subject>Queueing</subject><subject>Queuing</subject><subject>Resource allocation</subject><subject>Stochastic models</subject><subject>Supply & demand</subject><issn>0025-1909</issn><issn>1526-5501</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><recordid>eNqFkM1LwzAYxoMoOKdXzwHPrflsmuM2P2EqiIK3kKYpZGuTmXTK_ntbKnj09F5-z_O8_AC4xCjHpBTXnU8mJ4jQnEnKjsAMc1JknCN8DGYIEZ5hieQpOEtpgxASpShmYL1sna_hzcHrzhn4alPYR2Phom2D0b0LHjoPV21ItobPtv8OcZvgl9PwycUYIlxqs91Fm9I-2nNw0ug22YvfOwfvd7dvq4ds_XL_uFqsM0OY6DNcaoYJq7ipBZdYaFNLWaOGVlpKSrkwEhNijSmaihWEDlSlOa911dSNoIjOwdXUu4vhc29TrzbD136YVBRJTijnZTlQ-USZGFKKtlG76DodDwojNRpTozE1GlOjsSEAp4A1wbv0h5dSElQw9DEg2YQ434TYpf8qfwBFT3h5</recordid><startdate>20240801</startdate><enddate>20240801</enddate><creator>Kanoria, Yash</creator><general>INFORMS</general><general>Institute for Operations Research and the Management Sciences</general><scope>OQ6</scope><scope>AAYXX</scope><scope>CITATION</scope><scope>8BJ</scope><scope>FQK</scope><scope>JBE</scope><orcidid>https://orcid.org/0000-0002-1759-1009</orcidid><orcidid>https://orcid.org/0000-0002-7221-357X</orcidid></search><sort><creationdate>20240801</creationdate><title>Blind Dynamic Resource Allocation in Closed Networks via Mirror Backpressure</title><author>Kanoria, Yash</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c247t-18a4124b5cd75917acd99d0f3ba993357c9122ecc6fb4623d75ba55dabfdf7303</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Averages</topic><topic>backpressure</topic><topic>control of queueing networks</topic><topic>Customer services</topic><topic>Customers</topic><topic>Management science</topic><topic>mirror descent</topic><topic>no-underflow constraint</topic><topic>Operations management</topic><topic>Operations research</topic><topic>Optimization</topic><topic>Queueing</topic><topic>Queuing</topic><topic>Resource allocation</topic><topic>Stochastic models</topic><topic>Supply & demand</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Kanoria, Yash</creatorcontrib><collection>ECONIS</collection><collection>CrossRef</collection><collection>International Bibliography of the Social Sciences (IBSS)</collection><collection>International Bibliography of the Social Sciences</collection><collection>International Bibliography of the Social Sciences</collection><jtitle>Management science</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Kanoria, Yash</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Blind Dynamic Resource Allocation in Closed Networks via Mirror Backpressure</atitle><jtitle>Management science</jtitle><date>2024-08-01</date><risdate>2024</risdate><volume>70</volume><issue>8</issue><spage>5445</spage><epage>5462</epage><pages>5445-5462</pages><issn>0025-1909</issn><eissn>1526-5501</eissn><abstract>We study the problem of maximizing payoff generated over a period of time in a general class of closed queueing networks with a finite, fixed number of supply units that circulate in the system. Demand arrives stochastically, and serving a demand unit (customer) causes a supply unit to relocate from the “origin” to the “destination” of the customer. The key challenge is to manage the distribution of supply in the network. We consider general controls including customer entry control, pricing, and assignment. Motivating applications include shared transportation platforms and scrip systems. Inspired by the mirror descent algorithm for optimization and the backpressure policy for network control, we introduce a rich family of
mirror backpressure
(MBP) control policies. The MBP policies are simple and practical and crucially do not need any statistical knowledge of the demand (customer) arrival rates (these rates are permitted to vary in time). Under mild conditions, we propose MBP policies that are provably near optimal. Specifically, our policies lose at most
O
(
K
T
+
1
K
+
η
K
)
payoff per customer relative to the optimal policy that knows the demand arrival rates, where
K
is the number of supply units,
T
is the total number of customers over the time horizon, and
η
is the demand process’ average rate of change per customer arrival. An adaptation of MBP is found to perform well in numerical experiments based on data from NYC Cab.
This paper was accepted by Gabriel Weintraub, revenue management and market analytics.
Funding:
Y. Kanoria was supported by the National Science Foundation’s Division of Civil, Mechanical, and Manufacturing Innovation [Grant CMMI-1653477].
Supplemental Material:
The data files and online appendices are available at
https://doi.org/10.1287/mnsc.2023.4934
.</abstract><cop>Linthicum</cop><pub>INFORMS</pub><doi>10.1287/mnsc.2023.4934</doi><tpages>18</tpages><orcidid>https://orcid.org/0000-0002-1759-1009</orcidid><orcidid>https://orcid.org/0000-0002-7221-357X</orcidid></addata></record> |
fulltext | fulltext |
identifier | ISSN: 0025-1909 |
ispartof | Management science, 2024-08, Vol.70 (8), p.5445-5462 |
issn | 0025-1909 1526-5501 |
language | eng |
recordid | cdi_proquest_journals_3095235588 |
source | Informs |
subjects | Averages backpressure control of queueing networks Customer services Customers Management science mirror descent no-underflow constraint Operations management Operations research Optimization Queueing Queuing Resource allocation Stochastic models Supply & demand |
title | Blind Dynamic Resource Allocation in Closed Networks via Mirror Backpressure |
url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-24T21%3A41%3A38IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_cross&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Blind%20Dynamic%20Resource%20Allocation%20in%20Closed%20Networks%20via%20Mirror%20Backpressure&rft.jtitle=Management%20science&rft.au=Kanoria,%20Yash&rft.date=2024-08-01&rft.volume=70&rft.issue=8&rft.spage=5445&rft.epage=5462&rft.pages=5445-5462&rft.issn=0025-1909&rft.eissn=1526-5501&rft_id=info:doi/10.1287/mnsc.2023.4934&rft_dat=%3Cproquest_cross%3E3095235588%3C/proquest_cross%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=3095235588&rft_id=info:pmid/&rfr_iscdi=true |