20090306
Every optimization technique has inherent strengths and weaknesses. Moreover, some optimization algorithms contain characteristics which make them better suited to solve particular kinds of problems. Hybridization, or the combining of two or more complementary, but distinct methods, allows the user to take advantage of the beneficial elements of multiple methods. In this talk, we will describe an algorithm which combines statistical emulation via Gaussian process with pattern search optimization. We will demonstrate the applicability of our hybrid method to a problem of calibrating a computational model of an electrical circuit. In addition, we will describe how the treed Gaussian process can be used as a postprocessing tool to increase insight into the problem and discuss the usefulness of hybrid schemes for incorporating uncertainty.
Category: CE SeminarTechnische Universität Darmstadt
Graduate School CE
Dolivostraße 15
D64293 Darmstadt

Send email to assistants' office
Show a list of open BSc/MSc topics at GSC CE.