A comparison of RESTART implementations

Marnix J.J. Garvels; Dirk P. Kroese

A comparison of RESTART implementations

Marnix J.J. Garvels, Dirk P. Kroese

Stochastic Operations Research

Research output: Book/Report › Report › Other research output

51 Citations (Scopus)

Abstract

The RESTART method is a widely applicable simulation technique for the estimation of rare event probabilities. The method is based on the idea to restart the simulation in certain system states, in order to generate more occurrences of the rare event. One of the main questions for any RESTART implementation is {\em how} and {\em when} to restart the simulation, in order to achieve the most accurate results for a fixed simulation effort. In this paper we investigate and compare, both theoretically and empirically, different implementations of the RESTART method. We find that the original RESTART implementation, in which each path is split into a fixed number of copies, may not be the most efficient one. It is generally better to fix the total simulation effort for each stage of the simulation. Furthermore, given this effort, the best strategy is to restart an equal number of times from each state, rather than to restart each time from a randomly chosen state.

Original language	English
Place of Publication	Enschede
Publisher	Centre for Telematics and Information Technology (CTIT)
Number of pages	24
Publication status	Published - 1998

Publication series

Name	CTIT technical report series
Publisher	Centre for Telematics and Information Technology
No.	98-09
ISSN (Print)	1381-3625

51 Citations
1 Report
1 Conference contribution

A comparison of RESTART implementations
Garvels, M. J. J. & Kroese, D. P., 20 Feb 1998, Proceedings of the 1998 Winter Simulation Conference. Medeiros, D. J., Watson, E. F., Carson, J. S. & Manivannan, M. S. (eds.). Enschede: IEEE, Vol. 1. p. 601-609
Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

88 Downloads (Pure)
A comparison of RESTART implementations
Garvels, M. J. J. & Kroese, D. P., 1998, Enschede: University of Twente. 24 p. (Memorandum; no. 1443)
Research output: Book/Report › Report › Professional

Cite this

@book{127babfe89d847ae863f497f80a81da1,

title = "A comparison of RESTART implementations",

abstract = "The RESTART method is a widely applicable simulation technique for the estimation of rare event probabilities. The method is based on the idea to restart the simulation in certain system states, in order to generate more occurrences of the rare event. One of the main questions for any RESTART implementation is {\em how} and {\em when} to restart the simulation, in order to achieve the most accurate results for a fixed simulation effort. In this paper we investigate and compare, both theoretically and empirically, different implementations of the RESTART method. We find that the original RESTART implementation, in which each path is split into a fixed number of copies, may not be the most efficient one. It is generally better to fix the total simulation effort for each stage of the simulation. Furthermore, given this effort, the best strategy is to restart an equal number of times from each state, rather than to restart each time from a randomly chosen state.",

author = "Garvels, {Marnix J.J.} and Kroese, {Dirk P.}",

year = "1998",

language = "English",

series = "CTIT technical report series",

publisher = "Centre for Telematics and Information Technology (CTIT)",

number = "98-09",

address = "Netherlands",

}

TY - BOOK

T1 - A comparison of RESTART implementations

AU - Garvels, Marnix J.J.

AU - Kroese, Dirk P.

PY - 1998

Y1 - 1998

N2 - The RESTART method is a widely applicable simulation technique for the estimation of rare event probabilities. The method is based on the idea to restart the simulation in certain system states, in order to generate more occurrences of the rare event. One of the main questions for any RESTART implementation is {\em how} and {\em when} to restart the simulation, in order to achieve the most accurate results for a fixed simulation effort. In this paper we investigate and compare, both theoretically and empirically, different implementations of the RESTART method. We find that the original RESTART implementation, in which each path is split into a fixed number of copies, may not be the most efficient one. It is generally better to fix the total simulation effort for each stage of the simulation. Furthermore, given this effort, the best strategy is to restart an equal number of times from each state, rather than to restart each time from a randomly chosen state.

AB - The RESTART method is a widely applicable simulation technique for the estimation of rare event probabilities. The method is based on the idea to restart the simulation in certain system states, in order to generate more occurrences of the rare event. One of the main questions for any RESTART implementation is {\em how} and {\em when} to restart the simulation, in order to achieve the most accurate results for a fixed simulation effort. In this paper we investigate and compare, both theoretically and empirically, different implementations of the RESTART method. We find that the original RESTART implementation, in which each path is split into a fixed number of copies, may not be the most efficient one. It is generally better to fix the total simulation effort for each stage of the simulation. Furthermore, given this effort, the best strategy is to restart an equal number of times from each state, rather than to restart each time from a randomly chosen state.

M3 - Report

T3 - CTIT technical report series

BT - A comparison of RESTART implementations

PB - Centre for Telematics and Information Technology (CTIT)

CY - Enschede

ER -