Queuing with future information

Joel Spencer, Madhu Sudan, Kuang Xu

Research output: Contribution to journalArticle

Abstract

We study an admissions control problem, where a queue with service rate 1 - p receives incoming jobs at rate λ ∈ (1 - p, 1), and the decision maker is allowed to redirect away jobs up to a rate of p, with the objective of minimizing the time-Average queue length. We show that the amount of information about the future has a significant impact on system performance, in the heavy-Traffic regime. When the future is unknown, the optimal average queue length diverges at rate ~ log1/(1-p)1 1-λ, as →1. In sharp contrast, when all future arrival and service times are revealed beforehand, the optimal average queue length converges to a finite constant, (1 - p)/p, as →1. We further show that the finite limit of (1-p)/p can be achieve using only a finite lookahead window starting from the current time frame, whose length scales as O(log 1 1-λ ), as →1. This leads to the conjecture of an interesting duality between queuing delay and the amount of information about the future.

Original languageEnglish (US)
Pages (from-to)2091-2142
Number of pages52
JournalAnnals of Applied Probability
Volume24
Issue number5
DOIs
StatePublished - 2014

Fingerprint

Queuing
Queue Length
Heavy Traffic
Admission Control
Look-ahead
Time-average
Diverge
Length Scale
Queue
System Performance
Control Problem
Duality
Converge
Unknown

Keywords

  • Admissions control
  • Future information
  • Heavy-Traffic asymptotics
  • Offline
  • Online
  • Queuing theory
  • Random walk
  • Resource pooling

ASJC Scopus subject areas

  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Cite this

Queuing with future information. / Spencer, Joel; Sudan, Madhu; Xu, Kuang.

In: Annals of Applied Probability, Vol. 24, No. 5, 2014, p. 2091-2142.

Research output: Contribution to journalArticle

Spencer, Joel ; Sudan, Madhu ; Xu, Kuang. / Queuing with future information. In: Annals of Applied Probability. 2014 ; Vol. 24, No. 5. pp. 2091-2142.
@article{b300ec203c1e48cd9548ec9df1ea9c92,
title = "Queuing with future information",
abstract = "We study an admissions control problem, where a queue with service rate 1 - p receives incoming jobs at rate λ ∈ (1 - p, 1), and the decision maker is allowed to redirect away jobs up to a rate of p, with the objective of minimizing the time-Average queue length. We show that the amount of information about the future has a significant impact on system performance, in the heavy-Traffic regime. When the future is unknown, the optimal average queue length diverges at rate ~ log1/(1-p)1 1-λ, as →1. In sharp contrast, when all future arrival and service times are revealed beforehand, the optimal average queue length converges to a finite constant, (1 - p)/p, as →1. We further show that the finite limit of (1-p)/p can be achieve using only a finite lookahead window starting from the current time frame, whose length scales as O(log 1 1-λ ), as →1. This leads to the conjecture of an interesting duality between queuing delay and the amount of information about the future.",
keywords = "Admissions control, Future information, Heavy-Traffic asymptotics, Offline, Online, Queuing theory, Random walk, Resource pooling",
author = "Joel Spencer and Madhu Sudan and Kuang Xu",
year = "2014",
doi = "10.1214/13-AAP973",
language = "English (US)",
volume = "24",
pages = "2091--2142",
journal = "Annals of Applied Probability",
issn = "1050-5164",
publisher = "Institute of Mathematical Statistics",
number = "5",

}

TY - JOUR

T1 - Queuing with future information

AU - Spencer, Joel

AU - Sudan, Madhu

AU - Xu, Kuang

PY - 2014

Y1 - 2014

N2 - We study an admissions control problem, where a queue with service rate 1 - p receives incoming jobs at rate λ ∈ (1 - p, 1), and the decision maker is allowed to redirect away jobs up to a rate of p, with the objective of minimizing the time-Average queue length. We show that the amount of information about the future has a significant impact on system performance, in the heavy-Traffic regime. When the future is unknown, the optimal average queue length diverges at rate ~ log1/(1-p)1 1-λ, as →1. In sharp contrast, when all future arrival and service times are revealed beforehand, the optimal average queue length converges to a finite constant, (1 - p)/p, as →1. We further show that the finite limit of (1-p)/p can be achieve using only a finite lookahead window starting from the current time frame, whose length scales as O(log 1 1-λ ), as →1. This leads to the conjecture of an interesting duality between queuing delay and the amount of information about the future.

AB - We study an admissions control problem, where a queue with service rate 1 - p receives incoming jobs at rate λ ∈ (1 - p, 1), and the decision maker is allowed to redirect away jobs up to a rate of p, with the objective of minimizing the time-Average queue length. We show that the amount of information about the future has a significant impact on system performance, in the heavy-Traffic regime. When the future is unknown, the optimal average queue length diverges at rate ~ log1/(1-p)1 1-λ, as →1. In sharp contrast, when all future arrival and service times are revealed beforehand, the optimal average queue length converges to a finite constant, (1 - p)/p, as →1. We further show that the finite limit of (1-p)/p can be achieve using only a finite lookahead window starting from the current time frame, whose length scales as O(log 1 1-λ ), as →1. This leads to the conjecture of an interesting duality between queuing delay and the amount of information about the future.

KW - Admissions control

KW - Future information

KW - Heavy-Traffic asymptotics

KW - Offline

KW - Online

KW - Queuing theory

KW - Random walk

KW - Resource pooling

UR - http://www.scopus.com/inward/record.url?scp=84903975922&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84903975922&partnerID=8YFLogxK

U2 - 10.1214/13-AAP973

DO - 10.1214/13-AAP973

M3 - Article

AN - SCOPUS:84903975922

VL - 24

SP - 2091

EP - 2142

JO - Annals of Applied Probability

JF - Annals of Applied Probability

SN - 1050-5164

IS - 5

ER -