Optimal design of experiments for non‐linear response surface models

Yuanzhi Huang; Steven G. Gilmour; Kalliopi Mylona; Peter Goos

doi:10.1111/rssc.12313

Optimal design of experiments for non‐linear response surface models

Yuanzhi Huang, Steven G. Gilmour, Kalliopi Mylona, Peter Goos

Research output: Contribution to journal › Article › peer-review

19 Citations (Scopus)

190 Downloads (Pure)

Abstract

Many chemical and biological experiments involve multiple treatment factors and often it is convenient to fit a nonlinear model in these factors. This nonlinear model can be mechanistic, empirical or a hybrid of the two. Motivated by experiments in chemical engineering, we focus on D-optimal design for multifactor nonlinear response surfaces in general. In order to find and study optimal designs, we first implement conventional point and coordinate exchange algorithms. Next, we develop a novel multiphase optimisation method to construct D-optimal designs with improved properties. The benefits of this method are demonstrated by application to two experiments involving nonlinear regression models. The designs obtained are shown to be considerably more informative than designs obtained using traditional design optimality algorithms.

Original language	English
Pages (from-to)	623-640
Number of pages	18
Journal	APPLIED STATISTICS
Volume	68
Issue number	3
Early online date	7 Oct 2018
DOIs	https://doi.org/10.1111/rssc.12313
Publication status	Published - 27 Feb 2019

Keywords

Continuous optimization
D-optimality
Multifactor experiments
Multiphase optimization
Non-linear model
Parameter estimation

Access to Document

10.1111/rssc.12313

Optimal Design of Experiments_HUANG_Accepted28August2018_GREEN AAMAccepted author manuscript, 262 KB

Cite this

@article{349513f3c8664d109bcbeafac8bcf46b,

title = "Optimal design of experiments for non‐linear response surface models",

abstract = "Many chemical and biological experiments involve multiple treatment factors and often it is convenient to fit a nonlinear model in these factors. This nonlinear model can be mechanistic, empirical or a hybrid of the two. Motivated by experiments in chemical engineering, we focus on D-optimal design for multifactor nonlinear response surfaces in general. In order to find and study optimal designs, we first implement conventional point and coordinate exchange algorithms. Next, we develop a novel multiphase optimisation method to construct D-optimal designs with improved properties. The benefits of this method are demonstrated by application to two experiments involving nonlinear regression models. The designs obtained are shown to be considerably more informative than designs obtained using traditional design optimality algorithms.",

keywords = "Continuous optimization, D-optimality, Multifactor experiments, Multiphase optimization, Non-linear model, Parameter estimation",

author = "Yuanzhi Huang and Gilmour, {Steven G.} and Kalliopi Mylona and Peter Goos",

year = "2019",

month = feb,

day = "27",

doi = "10.1111/rssc.12313",

language = "English",

volume = "68",

pages = "623--640",

journal = "APPLIED STATISTICS",

issn = "0035-9254",

publisher = "Wiley-Blackwell",

number = "3",

}

TY - JOUR

T1 - Optimal design of experiments for non‐linear response surface models

AU - Huang, Yuanzhi

AU - Gilmour, Steven G.

AU - Mylona, Kalliopi

AU - Goos, Peter

PY - 2019/2/27

Y1 - 2019/2/27

N2 - Many chemical and biological experiments involve multiple treatment factors and often it is convenient to fit a nonlinear model in these factors. This nonlinear model can be mechanistic, empirical or a hybrid of the two. Motivated by experiments in chemical engineering, we focus on D-optimal design for multifactor nonlinear response surfaces in general. In order to find and study optimal designs, we first implement conventional point and coordinate exchange algorithms. Next, we develop a novel multiphase optimisation method to construct D-optimal designs with improved properties. The benefits of this method are demonstrated by application to two experiments involving nonlinear regression models. The designs obtained are shown to be considerably more informative than designs obtained using traditional design optimality algorithms.

AB - Many chemical and biological experiments involve multiple treatment factors and often it is convenient to fit a nonlinear model in these factors. This nonlinear model can be mechanistic, empirical or a hybrid of the two. Motivated by experiments in chemical engineering, we focus on D-optimal design for multifactor nonlinear response surfaces in general. In order to find and study optimal designs, we first implement conventional point and coordinate exchange algorithms. Next, we develop a novel multiphase optimisation method to construct D-optimal designs with improved properties. The benefits of this method are demonstrated by application to two experiments involving nonlinear regression models. The designs obtained are shown to be considerably more informative than designs obtained using traditional design optimality algorithms.

KW - Continuous optimization

KW - D-optimality

KW - Multifactor experiments

KW - Multiphase optimization

KW - Non-linear model

KW - Parameter estimation

UR - http://www.scopus.com/inward/record.url?scp=85054595581&partnerID=8YFLogxK

U2 - 10.1111/rssc.12313

DO - 10.1111/rssc.12313

M3 - Article

SN - 0035-9254

VL - 68

SP - 623

EP - 640

JO - APPLIED STATISTICS

JF - APPLIED STATISTICS

IS - 3

ER -

Optimal design of experiments for non‐linear response surface models

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this