Title
Derivative-Free Optimization of Expensive Functions with Computational Error Using Weighted Regression.
Abstract
We propose a derivative-free algorithm for optimizing computationally expensive functions with computational error. The algorithm is based on the trust region regression method by Conn, Scheinberg, and Vicente [A. R. Conn, K. Scheinberg, and L. N. Vicente, IMA J. Numer. Anal., 28 (2008), pp. 721-748] but uses weighted regression to obtain more accurate model functions at each trust region iteration. A heuristic weighting scheme is proposed that simultaneously handles (i) differing levels of uncertainty in function evaluations and (ii) errors induced by poor model fidelity. We also extend the theory of Lambda-poisedness and strong Lambda-poisedness to weighted regression. We report computational results comparing interpolation, regression, and weighted regression methods on a collection of benchmark problems. Weighted regression appears to outperform interpolation and regression models on nondifferentiable functions and functions with deterministic noise.
Year
DOI
Venue
2013
10.1137/100814688
SIAM JOURNAL ON OPTIMIZATION
Keywords
Field
DocType
derivative-free optimization,weighted regression models,noisy function evaluations
Trust region,Derivative-free optimization,Mathematical optimization,Weighting,Regression analysis,Polynomial regression,Unit-weighted regression,Computational mathematics,Interpolation,Mathematics
Journal
Volume
Issue
ISSN
23
1
1052-6234
Citations 
PageRank 
References 
9
0.62
10
Authors
3
Name
Order
Citations
PageRank
Stephen C. Billups120840.10
Jeffrey Larson2325.46
Peter Graf3122.36