Adjustable Robust Parameter Design with Unknown Distributions

Research output: Working paperDiscussion paperOther research output

361 Downloads (Pure)

Abstract

Abstract This article presents a novel combination of robust optimization developed in mathematical programming, and robust parameter design developed in statistical quality control. Robust parameter design uses metamodels estimated from experiments with both controllable and environmental inputs (factors). These experiments may be performed with either real or simulated systems; we focus on simulation experiments. For the environmental inputs, classic robust parameter design assumes known means and covariances, and sometimes even a known distribution. We, however, develop a robust optimization approach that uses only experimental data, so it does not need these classic assumptions. Moreover, we develop `adjustable' robust parameter design which adjusts the values of some or all of the controllable factors after observing the values of some or all of the environmental inputs. We also propose a new decision rule that is suitable for adjustable integer decision variables. We illustrate our novel method through several numerical examples, which demonstrate its effectiveness.
Original languageEnglish
Place of PublicationTilburg
PublisherEconometrics
Number of pages30
Volume2013-022
Publication statusPublished - 2013

Publication series

NameCentER Discussion Paper
Volume2013-022

Fingerprint

Experiments
Mathematical programming
Quality control

Keywords

  • robust optimization
  • simulation optimization
  • robust parameter design
  • phi-divergence

Cite this

Yanikoglu, I., den Hertog, D., & Kleijnen, J. P. C. (2013). Adjustable Robust Parameter Design with Unknown Distributions. (CentER Discussion Paper; Vol. 2013-022). Tilburg: Econometrics.
Yanikoglu, I. ; den Hertog, D. ; Kleijnen, Jack P.C. / Adjustable Robust Parameter Design with Unknown Distributions. Tilburg : Econometrics, 2013. (CentER Discussion Paper).
@techreport{47fec2281ffe48038e9758a9a822e96f,
title = "Adjustable Robust Parameter Design with Unknown Distributions",
abstract = "Abstract This article presents a novel combination of robust optimization developed in mathematical programming, and robust parameter design developed in statistical quality control. Robust parameter design uses metamodels estimated from experiments with both controllable and environmental inputs (factors). These experiments may be performed with either real or simulated systems; we focus on simulation experiments. For the environmental inputs, classic robust parameter design assumes known means and covariances, and sometimes even a known distribution. We, however, develop a robust optimization approach that uses only experimental data, so it does not need these classic assumptions. Moreover, we develop `adjustable' robust parameter design which adjusts the values of some or all of the controllable factors after observing the values of some or all of the environmental inputs. We also propose a new decision rule that is suitable for adjustable integer decision variables. We illustrate our novel method through several numerical examples, which demonstrate its effectiveness.",
keywords = "robust optimization, simulation optimization, robust parameter design, phi-divergence",
author = "I. Yanikoglu and {den Hertog}, D. and Kleijnen, {Jack P.C.}",
note = "Pagination: 30",
year = "2013",
language = "English",
volume = "2013-022",
series = "CentER Discussion Paper",
publisher = "Econometrics",
type = "WorkingPaper",
institution = "Econometrics",

}

Yanikoglu, I, den Hertog, D & Kleijnen, JPC 2013 'Adjustable Robust Parameter Design with Unknown Distributions' CentER Discussion Paper, vol. 2013-022, Econometrics, Tilburg.

Adjustable Robust Parameter Design with Unknown Distributions. / Yanikoglu, I.; den Hertog, D.; Kleijnen, Jack P.C.

Tilburg : Econometrics, 2013. (CentER Discussion Paper; Vol. 2013-022).

Research output: Working paperDiscussion paperOther research output

TY - UNPB

T1 - Adjustable Robust Parameter Design with Unknown Distributions

AU - Yanikoglu, I.

AU - den Hertog, D.

AU - Kleijnen, Jack P.C.

N1 - Pagination: 30

PY - 2013

Y1 - 2013

N2 - Abstract This article presents a novel combination of robust optimization developed in mathematical programming, and robust parameter design developed in statistical quality control. Robust parameter design uses metamodels estimated from experiments with both controllable and environmental inputs (factors). These experiments may be performed with either real or simulated systems; we focus on simulation experiments. For the environmental inputs, classic robust parameter design assumes known means and covariances, and sometimes even a known distribution. We, however, develop a robust optimization approach that uses only experimental data, so it does not need these classic assumptions. Moreover, we develop `adjustable' robust parameter design which adjusts the values of some or all of the controllable factors after observing the values of some or all of the environmental inputs. We also propose a new decision rule that is suitable for adjustable integer decision variables. We illustrate our novel method through several numerical examples, which demonstrate its effectiveness.

AB - Abstract This article presents a novel combination of robust optimization developed in mathematical programming, and robust parameter design developed in statistical quality control. Robust parameter design uses metamodels estimated from experiments with both controllable and environmental inputs (factors). These experiments may be performed with either real or simulated systems; we focus on simulation experiments. For the environmental inputs, classic robust parameter design assumes known means and covariances, and sometimes even a known distribution. We, however, develop a robust optimization approach that uses only experimental data, so it does not need these classic assumptions. Moreover, we develop `adjustable' robust parameter design which adjusts the values of some or all of the controllable factors after observing the values of some or all of the environmental inputs. We also propose a new decision rule that is suitable for adjustable integer decision variables. We illustrate our novel method through several numerical examples, which demonstrate its effectiveness.

KW - robust optimization

KW - simulation optimization

KW - robust parameter design

KW - phi-divergence

M3 - Discussion paper

VL - 2013-022

T3 - CentER Discussion Paper

BT - Adjustable Robust Parameter Design with Unknown Distributions

PB - Econometrics

CY - Tilburg

ER -

Yanikoglu I, den Hertog D, Kleijnen JPC. Adjustable Robust Parameter Design with Unknown Distributions. Tilburg: Econometrics. 2013. (CentER Discussion Paper).