Adaptive Importance Sampling Simulation of Queueing Networks

Pieter-Tjerk de Boer; V.F. Nicola; N. Rubinstein; Reuven Y. Rubinstein

doi:10.1109/WSC.2000.899776

Adaptive Importance Sampling Simulation of Queueing Networks

Pieter-Tjerk de Boer, V.F. Nicola, N. Rubinstein, Reuven Y. Rubinstein

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

35 Citations (Scopus)

226 Downloads (Pure)

Abstract

In this paper, a method is presented for the efficient estimation of rare-event (overflow) probabilities in Jackson queueing networks using importance sampling. The method differs in two ways from methods discussed in most earlier literature: the change of measure is state-dependent, i.e., it is a function of the content of the buffers, and the change of measure is determined using a cross-entropy-based adaptive procedure. This method yields asymptotically efficient estimation of overflow probabilities of queueing models for which it has been shown that methods using a stateindependent change of measure are not asymptotically efficient. Numerical results demonstrating the effectiveness of the method are presented as well.

Original language	Undefined
Title of host publication	The 2000 Winter Simulation Conference (WSC'00)
Place of Publication	Orlando, Florida, USA
Publisher	IEEE
Pages	646-655
Number of pages	10
ISBN (Print)	0-7803-6582-8
DOIs	https://doi.org/10.1109/WSC.2000.899776
Publication status	Published - Dec 2000
Event	2000 Winter Simulation Conference - Orlando, Florida, USA, Orlando, United States Duration: 10 Dec 2000 → 13 Dec 2000

Publication series

Name
Publisher	IEEE
Volume	1

Conference

Conference	2000 Winter Simulation Conference
Abbreviated title	WSC 2000
Country/Territory	United States
City	Orlando
Period	10/12/00 → 13/12/00
Other	10-13 December 2000

Keywords

EWI-7723
METIS-119564
IR-19042

Access to Document

10.1109/WSC.2000.899776

00899776

http://eprints.eemcs.utwente.nl/secure2/7723/01/086.PDF

Cite this

@inproceedings{c25b37772f2f41378dac783252a9da81,

title = "Adaptive Importance Sampling Simulation of Queueing Networks",

abstract = "In this paper, a method is presented for the efficient estimation of rare-event (overflow) probabilities in Jackson queueing networks using importance sampling. The method differs in two ways from methods discussed in most earlier literature: the change of measure is state-dependent, i.e., it is a function of the content of the buffers, and the change of measure is determined using a cross-entropy-based adaptive procedure. This method yields asymptotically efficient estimation of overflow probabilities of queueing models for which it has been shown that methods using a stateindependent change of measure are not asymptotically efficient. Numerical results demonstrating the effectiveness of the method are presented as well.",

keywords = "EWI-7723, METIS-119564, IR-19042",

author = "{de Boer}, Pieter-Tjerk and V.F. Nicola and N. Rubinstein and Rubinstein, {Reuven Y.}",

note = "Imported from research group DACS (ID number 284); 2000 Winter Simulation Conference ; Conference date: 10-12-2000 Through 13-12-2000",

year = "2000",

month = dec,

doi = "10.1109/WSC.2000.899776",

language = "Undefined",

isbn = "0-7803-6582-8",

publisher = "IEEE",

pages = "646--655",

booktitle = "The 2000 Winter Simulation Conference (WSC'00)",

address = "United States",

}

TY - GEN

T1 - Adaptive Importance Sampling Simulation of Queueing Networks

AU - de Boer, Pieter-Tjerk

AU - Nicola, V.F.

AU - Rubinstein, N.

AU - Rubinstein, Reuven Y.

N1 - Imported from research group DACS (ID number 284)

PY - 2000/12

Y1 - 2000/12

N2 - In this paper, a method is presented for the efficient estimation of rare-event (overflow) probabilities in Jackson queueing networks using importance sampling. The method differs in two ways from methods discussed in most earlier literature: the change of measure is state-dependent, i.e., it is a function of the content of the buffers, and the change of measure is determined using a cross-entropy-based adaptive procedure. This method yields asymptotically efficient estimation of overflow probabilities of queueing models for which it has been shown that methods using a stateindependent change of measure are not asymptotically efficient. Numerical results demonstrating the effectiveness of the method are presented as well.

AB - In this paper, a method is presented for the efficient estimation of rare-event (overflow) probabilities in Jackson queueing networks using importance sampling. The method differs in two ways from methods discussed in most earlier literature: the change of measure is state-dependent, i.e., it is a function of the content of the buffers, and the change of measure is determined using a cross-entropy-based adaptive procedure. This method yields asymptotically efficient estimation of overflow probabilities of queueing models for which it has been shown that methods using a stateindependent change of measure are not asymptotically efficient. Numerical results demonstrating the effectiveness of the method are presented as well.

KW - EWI-7723

KW - METIS-119564

KW - IR-19042

U2 - 10.1109/WSC.2000.899776

DO - 10.1109/WSC.2000.899776

M3 - Conference contribution

SN - 0-7803-6582-8

SP - 646

EP - 655

BT - The 2000 Winter Simulation Conference (WSC'00)

PB - IEEE

CY - Orlando, Florida, USA

T2 - 2000 Winter Simulation Conference

Y2 - 10 December 2000 through 13 December 2000

ER -