Throughput sensitivity analysis using a single simulation

Christoph Roser, Masaru Nakano, Minoru Tanaka

Research output: Chapter in Book/Report/Conference proceedingConference contribution

7 Citations (Scopus)

Abstract

This paper describes a novel method of calculating the sensitivity of the manufacturing system throughput to the variables of the machines. The sensitivity analysis needs only a single simulation, yet is easy to use and provides accurate results. This sensitivity analysis is then used to predict the change in the system throughput due to a change of the variables of the machines provided that the system change does not significantly change the bottleneck. These predictions can be used for a local optimization, allowing the use of a steepest descent optimization algorithm. The method is based on improving the momentary shifting bottlenecks. The shifting bottlenecks are detected using the shifting bottleneck detection method based on the active duration, i.e., the time a machine is active without interruption. The method is easy to understand and easy to implement in existing simulation software.

Original languageEnglish
Title of host publicationWinter Simulation Conference Proceedings
EditorsE. Yucesan, C.H. Chen, J.L. Snowdon, J.M. Charnes
Pages1087-1094
Number of pages8
Volume2
Publication statusPublished - 2002
Externally publishedYes
EventProceedings of the 2002 Winter Simulation Conference - San Diego, CA, United States
Duration: 2002 Dec 82002 Dec 11

Other

OtherProceedings of the 2002 Winter Simulation Conference
CountryUnited States
CitySan Diego, CA
Period02/12/802/12/11

Fingerprint

Sensitivity analysis
Sensitivity Analysis
Throughput
Simulation
Local Optimization
Descent Algorithm
Steepest Descent
Simulation Software
Optimization Algorithm
Predict
Prediction

ASJC Scopus subject areas

  • Chemical Health and Safety
  • Software
  • Safety, Risk, Reliability and Quality
  • Applied Mathematics
  • Modelling and Simulation

Cite this

Roser, C., Nakano, M., & Tanaka, M. (2002). Throughput sensitivity analysis using a single simulation. In E. Yucesan, C. H. Chen, J. L. Snowdon, & J. M. Charnes (Eds.), Winter Simulation Conference Proceedings (Vol. 2, pp. 1087-1094)

Throughput sensitivity analysis using a single simulation. / Roser, Christoph; Nakano, Masaru; Tanaka, Minoru.

Winter Simulation Conference Proceedings. ed. / E. Yucesan; C.H. Chen; J.L. Snowdon; J.M. Charnes. Vol. 2 2002. p. 1087-1094.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Roser, C, Nakano, M & Tanaka, M 2002, Throughput sensitivity analysis using a single simulation. in E Yucesan, CH Chen, JL Snowdon & JM Charnes (eds), Winter Simulation Conference Proceedings. vol. 2, pp. 1087-1094, Proceedings of the 2002 Winter Simulation Conference, San Diego, CA, United States, 02/12/8.
Roser C, Nakano M, Tanaka M. Throughput sensitivity analysis using a single simulation. In Yucesan E, Chen CH, Snowdon JL, Charnes JM, editors, Winter Simulation Conference Proceedings. Vol. 2. 2002. p. 1087-1094
Roser, Christoph ; Nakano, Masaru ; Tanaka, Minoru. / Throughput sensitivity analysis using a single simulation. Winter Simulation Conference Proceedings. editor / E. Yucesan ; C.H. Chen ; J.L. Snowdon ; J.M. Charnes. Vol. 2 2002. pp. 1087-1094
@inproceedings{21887e4af47144c29c5ef85f60dfa129,
title = "Throughput sensitivity analysis using a single simulation",
abstract = "This paper describes a novel method of calculating the sensitivity of the manufacturing system throughput to the variables of the machines. The sensitivity analysis needs only a single simulation, yet is easy to use and provides accurate results. This sensitivity analysis is then used to predict the change in the system throughput due to a change of the variables of the machines provided that the system change does not significantly change the bottleneck. These predictions can be used for a local optimization, allowing the use of a steepest descent optimization algorithm. The method is based on improving the momentary shifting bottlenecks. The shifting bottlenecks are detected using the shifting bottleneck detection method based on the active duration, i.e., the time a machine is active without interruption. The method is easy to understand and easy to implement in existing simulation software.",
author = "Christoph Roser and Masaru Nakano and Minoru Tanaka",
year = "2002",
language = "English",
volume = "2",
pages = "1087--1094",
editor = "E. Yucesan and C.H. Chen and J.L. Snowdon and J.M. Charnes",
booktitle = "Winter Simulation Conference Proceedings",

}

TY - GEN

T1 - Throughput sensitivity analysis using a single simulation

AU - Roser, Christoph

AU - Nakano, Masaru

AU - Tanaka, Minoru

PY - 2002

Y1 - 2002

N2 - This paper describes a novel method of calculating the sensitivity of the manufacturing system throughput to the variables of the machines. The sensitivity analysis needs only a single simulation, yet is easy to use and provides accurate results. This sensitivity analysis is then used to predict the change in the system throughput due to a change of the variables of the machines provided that the system change does not significantly change the bottleneck. These predictions can be used for a local optimization, allowing the use of a steepest descent optimization algorithm. The method is based on improving the momentary shifting bottlenecks. The shifting bottlenecks are detected using the shifting bottleneck detection method based on the active duration, i.e., the time a machine is active without interruption. The method is easy to understand and easy to implement in existing simulation software.

AB - This paper describes a novel method of calculating the sensitivity of the manufacturing system throughput to the variables of the machines. The sensitivity analysis needs only a single simulation, yet is easy to use and provides accurate results. This sensitivity analysis is then used to predict the change in the system throughput due to a change of the variables of the machines provided that the system change does not significantly change the bottleneck. These predictions can be used for a local optimization, allowing the use of a steepest descent optimization algorithm. The method is based on improving the momentary shifting bottlenecks. The shifting bottlenecks are detected using the shifting bottleneck detection method based on the active duration, i.e., the time a machine is active without interruption. The method is easy to understand and easy to implement in existing simulation software.

UR - http://www.scopus.com/inward/record.url?scp=0036932018&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0036932018&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:0036932018

VL - 2

SP - 1087

EP - 1094

BT - Winter Simulation Conference Proceedings

A2 - Yucesan, E.

A2 - Chen, C.H.

A2 - Snowdon, J.L.

A2 - Charnes, J.M.

ER -