Title :
Dealing with Model Errors in Approximation Model-Based Optimization
Author :
Wang, Linda ; Lowther, David A.
Author_Institution :
Dept. of Electr. & Comput. Eng., McGill Univ., Montreal, Que.
Abstract :
Approximation models are often used in place of complex analysis code during optimization. Uncertainties in the model parameters could lead to inaccuracies in the process. This paper presents a Bayesian approach for finding an optimum that is robust to the uncertainty in model parameters. We test the proposed methodology with two model types applied to standard electromagnetics benchmark problems
Keywords :
Bayes methods; approximation theory; electromagnetic devices; optimisation; Bayesian approach; approximation model; optimization; standard electromagnetics benchmark problems; uncertainty parameters; Bayesian methods; Computer errors; Design optimization; Parameter estimation; Polynomials; Predictive models; Robustness; Testing; Uncertain systems; Uncertainty;
Conference_Titel :
Electromagnetic Field Computation, 2006 12th Biennial IEEE Conference on
Conference_Location :
Miami, FL
Print_ISBN :
1-4244-0320-0
DOI :
10.1109/CEFC-06.2006.1633022
