Efficient multi-fidelity simulation optimization

Jie Xu, Si Zhang, Edward Huang, Chun Hung Chen, Loo Hay Lee, Nurcin Celik

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

22 Citations (Scopus)

Abstract

Simulation models of different fidelity levels are often available for a complex system. High-fidelity simulations are accurate but time-consuming. Therefore, they can only be applied to a small number of solutions. Low-fidelity simulations are faster and can evaluate a large number of solutions. But their results may contain significant bias and variability. We propose an Multi-fidelity Optimization with Ordinal Transformation and Optimal Sampling (MO<sup>2</sup>TOS) framework to exploit the benefits of high- and low-fidelity simulations to efficiently identify a (near) optimal solution. MO2TOS uses low-fidelity simulations for all solutions and then assigns a fixed budget of high-fidelity simulations to solutions based on low-fidelity simulation results. We show the benefits of MO<sup>2</sup>TOS via theoretical analysis and numerical experiments with deterministic simulations and stochastic simulations where noise is negligible with sufficient replications. We compare MO<sup>2</sup>TOS to Equal Allocation (EA) and Optimal Computing Budget Allocation (OCBA). MO<sup>2</sup>TOS consistently outperforms both EA and OCBA.

Original languageEnglish (US)
Title of host publicationProceedings - Winter Simulation Conference
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages3940-3951
Number of pages12
Volume2015-January
ISBN (Print)9781479974863
DOIs
StatePublished - Jan 23 2015
Event2014 Winter Simulation Conference, WSC 2014 - Savannah, United States
Duration: Dec 7 2014Dec 10 2014

Other

Other2014 Winter Simulation Conference, WSC 2014
CountryUnited States
CitySavannah
Period12/7/1412/10/14

Fingerprint

Simulation Optimization
Fidelity
Simulation
Number of Solutions
Large scale systems
Sampling
Computing
Stochastic Simulation
Replication
Assign
Experiments
Complex Systems
Theoretical Analysis
Simulation Model
Optimal Solution
Numerical Experiment
Sufficient
Optimization
Evaluate

ASJC Scopus subject areas

  • Software
  • Modeling and Simulation
  • Computer Science Applications

Cite this

Xu, J., Zhang, S., Huang, E., Chen, C. H., Lee, L. H., & Celik, N. (2015). Efficient multi-fidelity simulation optimization. In Proceedings - Winter Simulation Conference (Vol. 2015-January, pp. 3940-3951). [7020219] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/WSC.2014.7020219

Efficient multi-fidelity simulation optimization. / Xu, Jie; Zhang, Si; Huang, Edward; Chen, Chun Hung; Lee, Loo Hay; Celik, Nurcin.

Proceedings - Winter Simulation Conference. Vol. 2015-January Institute of Electrical and Electronics Engineers Inc., 2015. p. 3940-3951 7020219.

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

Xu, J, Zhang, S, Huang, E, Chen, CH, Lee, LH & Celik, N 2015, Efficient multi-fidelity simulation optimization. in Proceedings - Winter Simulation Conference. vol. 2015-January, 7020219, Institute of Electrical and Electronics Engineers Inc., pp. 3940-3951, 2014 Winter Simulation Conference, WSC 2014, Savannah, United States, 12/7/14. https://doi.org/10.1109/WSC.2014.7020219
Xu J, Zhang S, Huang E, Chen CH, Lee LH, Celik N. Efficient multi-fidelity simulation optimization. In Proceedings - Winter Simulation Conference. Vol. 2015-January. Institute of Electrical and Electronics Engineers Inc. 2015. p. 3940-3951. 7020219 https://doi.org/10.1109/WSC.2014.7020219
Xu, Jie ; Zhang, Si ; Huang, Edward ; Chen, Chun Hung ; Lee, Loo Hay ; Celik, Nurcin. / Efficient multi-fidelity simulation optimization. Proceedings - Winter Simulation Conference. Vol. 2015-January Institute of Electrical and Electronics Engineers Inc., 2015. pp. 3940-3951
@inproceedings{faa57b6b26e44d7fbd49bba332a28ebc,
title = "Efficient multi-fidelity simulation optimization",
abstract = "Simulation models of different fidelity levels are often available for a complex system. High-fidelity simulations are accurate but time-consuming. Therefore, they can only be applied to a small number of solutions. Low-fidelity simulations are faster and can evaluate a large number of solutions. But their results may contain significant bias and variability. We propose an Multi-fidelity Optimization with Ordinal Transformation and Optimal Sampling (MO2TOS) framework to exploit the benefits of high- and low-fidelity simulations to efficiently identify a (near) optimal solution. MO2TOS uses low-fidelity simulations for all solutions and then assigns a fixed budget of high-fidelity simulations to solutions based on low-fidelity simulation results. We show the benefits of MO2TOS via theoretical analysis and numerical experiments with deterministic simulations and stochastic simulations where noise is negligible with sufficient replications. We compare MO2TOS to Equal Allocation (EA) and Optimal Computing Budget Allocation (OCBA). MO2TOS consistently outperforms both EA and OCBA.",
author = "Jie Xu and Si Zhang and Edward Huang and Chen, {Chun Hung} and Lee, {Loo Hay} and Nurcin Celik",
year = "2015",
month = "1",
day = "23",
doi = "10.1109/WSC.2014.7020219",
language = "English (US)",
isbn = "9781479974863",
volume = "2015-January",
pages = "3940--3951",
booktitle = "Proceedings - Winter Simulation Conference",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - Efficient multi-fidelity simulation optimization

AU - Xu, Jie

AU - Zhang, Si

AU - Huang, Edward

AU - Chen, Chun Hung

AU - Lee, Loo Hay

AU - Celik, Nurcin

PY - 2015/1/23

Y1 - 2015/1/23

N2 - Simulation models of different fidelity levels are often available for a complex system. High-fidelity simulations are accurate but time-consuming. Therefore, they can only be applied to a small number of solutions. Low-fidelity simulations are faster and can evaluate a large number of solutions. But their results may contain significant bias and variability. We propose an Multi-fidelity Optimization with Ordinal Transformation and Optimal Sampling (MO2TOS) framework to exploit the benefits of high- and low-fidelity simulations to efficiently identify a (near) optimal solution. MO2TOS uses low-fidelity simulations for all solutions and then assigns a fixed budget of high-fidelity simulations to solutions based on low-fidelity simulation results. We show the benefits of MO2TOS via theoretical analysis and numerical experiments with deterministic simulations and stochastic simulations where noise is negligible with sufficient replications. We compare MO2TOS to Equal Allocation (EA) and Optimal Computing Budget Allocation (OCBA). MO2TOS consistently outperforms both EA and OCBA.

AB - Simulation models of different fidelity levels are often available for a complex system. High-fidelity simulations are accurate but time-consuming. Therefore, they can only be applied to a small number of solutions. Low-fidelity simulations are faster and can evaluate a large number of solutions. But their results may contain significant bias and variability. We propose an Multi-fidelity Optimization with Ordinal Transformation and Optimal Sampling (MO2TOS) framework to exploit the benefits of high- and low-fidelity simulations to efficiently identify a (near) optimal solution. MO2TOS uses low-fidelity simulations for all solutions and then assigns a fixed budget of high-fidelity simulations to solutions based on low-fidelity simulation results. We show the benefits of MO2TOS via theoretical analysis and numerical experiments with deterministic simulations and stochastic simulations where noise is negligible with sufficient replications. We compare MO2TOS to Equal Allocation (EA) and Optimal Computing Budget Allocation (OCBA). MO2TOS consistently outperforms both EA and OCBA.

UR - http://www.scopus.com/inward/record.url?scp=84940538573&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84940538573&partnerID=8YFLogxK

U2 - 10.1109/WSC.2014.7020219

DO - 10.1109/WSC.2014.7020219

M3 - Conference contribution

SN - 9781479974863

VL - 2015-January

SP - 3940

EP - 3951

BT - Proceedings - Winter Simulation Conference

PB - Institute of Electrical and Electronics Engineers Inc.

ER -