Global and local optimization using radial basis function response surface models

Abstract

The focus of this paper is the optimization of complex multi-parameter systems. We consider systems in which the objective function is not known explicitly, and can only be evaluated through computationally intensive numerical simulation or through costly physical experiments. The objective function may also contain many local extrema which may be of interest. Given objective function values at a scattered set of parameter values, we develop a response surface model that can dramatically reduce the required computation time for parameter optimization runs. The response surface model is developed using radial basis functions, producing a model whose objective function values match those of the original system at all sampled data points. Interpolation to any other point is easily accomplished and generates a model which represents the system over the entire parameter space. This paper presents the details of the use of radial basis functions to transform scattered data points, obtained from a complex continuum mechanics simulation of explosive materials, into a response surface model of a function over the given parameter space. Response surface methodology and radial basis functions are discussed in general and are applied to a global optimization problem for an explosive oil well penetrator.

Effective design of semiconductor devices using evolutionary-based derivative free optimization

“…Finally, due to increasing computational effort required to simulate these devices, we are introducing surrogate optimization approaches based on DACE [22], Nadaraya-Watson estimator [23] and Radial Basis Functions. [24].…”

Modelos De Malla Basados en Estimadores (A,u,?)

“…Resumiendo lo explicado por Baxter (1992), Mcdonalda (2007) y Sánchez--Torres y Branch (2009) este método se basa en suponer que para el número real R≥0, denominado radio, la estimación U e en el punto P e se obtiene mediante la ecuación siguiente:…”

Computational Reduction for Parametrized PDEs: Strategies and Applications

“…In the case where the output function s(·) has several local maxima and minima, and the problem involves high dimensions and/or scattered data in the parameter space, radial basis functions (RBF) have been found very accurate to generate response surface models [33]. An interpolation model based on RBFs is a linear combination of the form…”