Computer-aided methodologies for optimizing complex problems
Dr. Matteo Borrotti, IMATI-CNR, Milan
Sala conferenze IMATI-CNR - Thursday, October 8, 2015 h.16:00
Abstract. In recent years, the use of rules that govern experiments for technological
improvement (i.e., Design of Experiments) has received renewed momentum
through the utilization of Computer Experiments. These experiments are run
on a computer code implementing a simulation model of a physical system of
interest such as in the process of recovering waste printed circuit boards
(PCBs) or in enzyme kinetics or in prosthesis design. In general, a
simulation model, or simulator, consists of a set of many linear or
nonlinear, ordinary and/or differential simultaneous equations, whose
solutions may not be amenable to analytical expression. Furthermore, runs
of the simulator can be expensive and/or time-consuming. In this case, the
use of a surrogate model (metamodel) is suggested. These metamodels are
simpler models, which represent a valid approximation of the original
simulator and usually are statistical interpolators built from the
simulated input-output data. Predictions at untried experimental points can
then be made by metamodels, and optimization techniques can be applied in
order to find best settings.
The aim of this presentation is to give an overview of computer experiments
techniques through the description of main concepts and meaningful examples.
Università degli Studi di Pavia -
Via Ferrata, 5 - 27100 Pavia
Tel +39.0382.985600 - Fax +39.0382.985602