Abstract

Engineering designs often involve finding solutions to complex design optimization problems. Other examples of real world problems that involve complex optimization are large-scale finite element analysis (FEA), computational fluid dynamics (CFD), etc. Stochastic, zero-order search techniques like evolutionary algorithms (EAs) are efficient global optimizers. However, these methods require numerous evaluations of candidate solutions. For the real world problems, like engineering design optimization, large scale finite element analysis (FEA), such evaluation of one single function could involve hours to days of complete simulation. The solution to this problem lies in reducing the number of such evaluations without significantly compromising the accuracy of the results. This chapter describes how meta modelling can be used to achieve this, with an example of the DAFHEA-II (enhanced Dynamic Approximate Fitness based Hybrid Evolutionary Algorithm) technique.
Original languageEnglish
Title of host publicationDAAAM International Scienticic Book
Place of PublicationVienna, Austria
PublisherDAAAM International Publishing
Pages015-030
Number of pages16
Edition3
ISBN (Print)390150947X
Publication statusPublished - 2006

Fingerprint

Dive into the research topics of 'Meta-modelling for EA based Design Optimization'. Together they form a unique fingerprint.

Cite this