Identification of response surface models using genetic programming

TL Lew, AB Spencer, FL Scarpa, K Worden, A Rutherford, F Hemez

Research output: Contribution to journalArticle (Academic Journal)

28 Citations (Scopus)

Abstract

There is a move in modern research in Structural Dynamics towards analysing the inherent uncertainty in a given problem. This may be quantifying or fusing uncertainty models, or can be propagation of uncertainty through a system or calculation. If the system of interest is represented by, e.g. a large Finite Element (FE) model the large number of computations involved can rule out many approaches due to the expense of carrying out many runs. One way of circumnavigating this problem is to replace the true system by an approximate surrogate/replacement model, which is fast-running compared to the original. In traditional approaches using response surfaces a simple least-squares multinomial model is often adopted. The objective of this paper is to extend the class of possible models considerably by carrying out a general symbolic regression using a Genetic Programming approach. The approach is demonstrated on both univariate and multivariate problems with both computational and experimental data.
Translated title of the contributionIdentification of response surface models using genetic programming
Original languageEnglish
Pages (from-to)1819 - 1831
Number of pages13
JournalMechanical Systems and Signal Processing
Volume20
DOIs
Publication statusPublished - Aug 2006

Bibliographical note

Author of Publication Reviewed: Lew TL , Spencer AB , Scarpa F , Worden K , Rutherford A , Hemez F
Publisher: ACADEMIC PRESS LTD ELSEVIER SCIENCE LTD, 24-28 OVAL RD, LONDON NW1 7DX, ENGLAND
Other identifier: IDS Number: 084SZ

Fingerprint Dive into the research topics of 'Identification of response surface models using genetic programming'. Together they form a unique fingerprint.

Cite this