2 Citations (Scopus)

Abstract

Analytical and simulation models are two common types of approaches used to estimate and predict the performance of complex production systems. Typically analytical models are fast to run but can have reduced accuracy. On the other hand simulation models can achieve high accuracy, but only at the cost of large simulation time and number of replications. Traditionally, the research has been focusing on the development of models able to achieve a satisfactory trade off between accuracy and computational effort. Nevertheless, such an approach implies the choice of a single model to approximate the system behavior. There is still lack of a generic model that can deliver high accuracy and low computational cost for production systems. In this paper, we attempt to address this issue and present a multi-fidelity modeling approach, utilizing both analytical models and simulation models at different levels of fidelity, to efficiently and effectively estimate the performance of asynchronous serial lines with exponential machines. Experimental results show that the multi-fidelity model provides better estimation of the production rate of the studied example Such a model has demonstrated potential in evaluating a large number of solutions with limited computational budget.

Original languageEnglish (US)
Title of host publication2017 13th IEEE Conference on Automation Science and Engineering, CASE 2017
PublisherIEEE Computer Society
Pages30-35
Number of pages6
Volume2017-August
ISBN (Electronic)9781509067800
DOIs
StatePublished - Jan 12 2018
Event13th IEEE Conference on Automation Science and Engineering, CASE 2017 - Xi'an, China
Duration: Aug 20 2017Aug 23 2017

Other

Other13th IEEE Conference on Automation Science and Engineering, CASE 2017
CountryChina
CityXi'an
Period8/20/178/23/17

Fingerprint

Analytical models
Costs

Keywords

  • analytical modeling
  • Multi-fidelity modeling
  • serial production line
  • simulation

ASJC Scopus subject areas

  • Control and Systems Engineering
  • Electrical and Electronic Engineering

Cite this

Kang, Y., Mathesen, L., Pedrielli, G., & Ju, F. (2018). Multi-fidelity modeling for analysis of serial production lines. In 2017 13th IEEE Conference on Automation Science and Engineering, CASE 2017 (Vol. 2017-August, pp. 30-35). IEEE Computer Society. https://doi.org/10.1109/COASE.2017.8256071

Multi-fidelity modeling for analysis of serial production lines. / Kang, Yunyi; Mathesen, Logan; Pedrielli, Giulia; Ju, Feng.

2017 13th IEEE Conference on Automation Science and Engineering, CASE 2017. Vol. 2017-August IEEE Computer Society, 2018. p. 30-35.

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

Kang, Y, Mathesen, L, Pedrielli, G & Ju, F 2018, Multi-fidelity modeling for analysis of serial production lines. in 2017 13th IEEE Conference on Automation Science and Engineering, CASE 2017. vol. 2017-August, IEEE Computer Society, pp. 30-35, 13th IEEE Conference on Automation Science and Engineering, CASE 2017, Xi'an, China, 8/20/17. https://doi.org/10.1109/COASE.2017.8256071
Kang Y, Mathesen L, Pedrielli G, Ju F. Multi-fidelity modeling for analysis of serial production lines. In 2017 13th IEEE Conference on Automation Science and Engineering, CASE 2017. Vol. 2017-August. IEEE Computer Society. 2018. p. 30-35 https://doi.org/10.1109/COASE.2017.8256071
Kang, Yunyi ; Mathesen, Logan ; Pedrielli, Giulia ; Ju, Feng. / Multi-fidelity modeling for analysis of serial production lines. 2017 13th IEEE Conference on Automation Science and Engineering, CASE 2017. Vol. 2017-August IEEE Computer Society, 2018. pp. 30-35
@inproceedings{ffea30792c294cd49b151f6f407977cc,
title = "Multi-fidelity modeling for analysis of serial production lines",
abstract = "Analytical and simulation models are two common types of approaches used to estimate and predict the performance of complex production systems. Typically analytical models are fast to run but can have reduced accuracy. On the other hand simulation models can achieve high accuracy, but only at the cost of large simulation time and number of replications. Traditionally, the research has been focusing on the development of models able to achieve a satisfactory trade off between accuracy and computational effort. Nevertheless, such an approach implies the choice of a single model to approximate the system behavior. There is still lack of a generic model that can deliver high accuracy and low computational cost for production systems. In this paper, we attempt to address this issue and present a multi-fidelity modeling approach, utilizing both analytical models and simulation models at different levels of fidelity, to efficiently and effectively estimate the performance of asynchronous serial lines with exponential machines. Experimental results show that the multi-fidelity model provides better estimation of the production rate of the studied example Such a model has demonstrated potential in evaluating a large number of solutions with limited computational budget.",
keywords = "analytical modeling, Multi-fidelity modeling, serial production line, simulation",
author = "Yunyi Kang and Logan Mathesen and Giulia Pedrielli and Feng Ju",
year = "2018",
month = "1",
day = "12",
doi = "10.1109/COASE.2017.8256071",
language = "English (US)",
volume = "2017-August",
pages = "30--35",
booktitle = "2017 13th IEEE Conference on Automation Science and Engineering, CASE 2017",
publisher = "IEEE Computer Society",

}

TY - GEN

T1 - Multi-fidelity modeling for analysis of serial production lines

AU - Kang, Yunyi

AU - Mathesen, Logan

AU - Pedrielli, Giulia

AU - Ju, Feng

PY - 2018/1/12

Y1 - 2018/1/12

N2 - Analytical and simulation models are two common types of approaches used to estimate and predict the performance of complex production systems. Typically analytical models are fast to run but can have reduced accuracy. On the other hand simulation models can achieve high accuracy, but only at the cost of large simulation time and number of replications. Traditionally, the research has been focusing on the development of models able to achieve a satisfactory trade off between accuracy and computational effort. Nevertheless, such an approach implies the choice of a single model to approximate the system behavior. There is still lack of a generic model that can deliver high accuracy and low computational cost for production systems. In this paper, we attempt to address this issue and present a multi-fidelity modeling approach, utilizing both analytical models and simulation models at different levels of fidelity, to efficiently and effectively estimate the performance of asynchronous serial lines with exponential machines. Experimental results show that the multi-fidelity model provides better estimation of the production rate of the studied example Such a model has demonstrated potential in evaluating a large number of solutions with limited computational budget.

AB - Analytical and simulation models are two common types of approaches used to estimate and predict the performance of complex production systems. Typically analytical models are fast to run but can have reduced accuracy. On the other hand simulation models can achieve high accuracy, but only at the cost of large simulation time and number of replications. Traditionally, the research has been focusing on the development of models able to achieve a satisfactory trade off between accuracy and computational effort. Nevertheless, such an approach implies the choice of a single model to approximate the system behavior. There is still lack of a generic model that can deliver high accuracy and low computational cost for production systems. In this paper, we attempt to address this issue and present a multi-fidelity modeling approach, utilizing both analytical models and simulation models at different levels of fidelity, to efficiently and effectively estimate the performance of asynchronous serial lines with exponential machines. Experimental results show that the multi-fidelity model provides better estimation of the production rate of the studied example Such a model has demonstrated potential in evaluating a large number of solutions with limited computational budget.

KW - analytical modeling

KW - Multi-fidelity modeling

KW - serial production line

KW - simulation

UR - http://www.scopus.com/inward/record.url?scp=85044966358&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85044966358&partnerID=8YFLogxK

U2 - 10.1109/COASE.2017.8256071

DO - 10.1109/COASE.2017.8256071

M3 - Conference contribution

AN - SCOPUS:85044966358

VL - 2017-August

SP - 30

EP - 35

BT - 2017 13th IEEE Conference on Automation Science and Engineering, CASE 2017

PB - IEEE Computer Society

ER -