Library Subscription: Guest
Begell Digital Portal Begell Digital Library eBooks Journals References & Proceedings Research Collections
International Journal for Uncertainty Quantification
IF: 4.911 5-Year IF: 3.179 SJR: 1.008 SNIP: 0.983 CiteScore™: 5.2

ISSN Print: 2152-5080
ISSN Online: 2152-5099

Open Access

International Journal for Uncertainty Quantification

DOI: 10.1615/Int.J.UncertaintyQuantification.2014006730
pages 479-510

GRADIENT-BASED STOCHASTIC OPTIMIZATION METHODS IN BAYESIAN EXPERIMENTAL DESIGN

Xun Huan
Sandia National Laboratories, 7011 East Ave, MS 9051, Livermore, CA 94550, USA; Department of Mechanical Engineering, University of Michigan, Ann Arbor, MI 48109, USA
Youssef Marzouk
Department of Aeronautics and Astronautics, Massachusetts Institute of Technology, 77 Massachusetts Ave, Room 33-305 Cambridge, MA 02139 USA

ABSTRACT

Optimal experimental design (OED) seeks experiments expected to yield the most useful data for some purpose. In practical circumstances where experiments are time-consuming or resource-intensive, OED can yield enormous savings. We pursue OED for nonlinear systems from a Bayesian perspective, with the goal of choosing experiments that are optimal for parameter inference. Our objective in this context is the expected information gain in model parameters, which in general can only be estimated using Monte Carlo methods. Maximizing this objective thus becomes a stochastic optimization problem. This paper develops gradient-based stochastic optimization methods for the design of experiments on a continuous parameter space. Given a Monte Carlo estimator of expected information gain, we use infinitesimal perturbation analysis to derive gradients of this estimator.We are then able to formulate two gradient-based stochastic optimization approaches: (i) Robbins-Monro stochastic approximation, and (ii) sample average approximation combined with a deterministic quasi-Newton method. A polynomial chaos approximation of the forward model accelerates objective and gradient evaluations in both cases.We discuss the implementation of these optimization methods, then conduct an empirical comparison of their performance. To demonstrate design in a nonlinear setting with partial differential equation forward models, we use the problem of sensor placement for source inversion. Numerical results yield useful guidelines on the choice of algorithm and sample sizes, assess the impact of estimator bias, and quantify tradeoffs of computational cost versus solution quality and robustness.


Articles with similar content:

EMPIRICAL EVALUATION OF BAYESIAN OPTIMIZATION IN PARAMETRIC TUNING OF CHAOTIC SYSTEMS
International Journal for Uncertainty Quantification, Vol.6, 2016, issue 6
Antti Solonen, Heikki Jarvinen, Janne Hakkarainen, Mudassar Abbas, Erkki Oja, Alexander Ilin
COMPUTING GREEN'S FUNCTIONS FOR FLOW IN HETEROGENEOUS COMPOSITE MEDIA
International Journal for Uncertainty Quantification, Vol.3, 2013, issue 1
David A. Barajas-Solano, Daniel M. Tartakovsky
STOCHASTIC MULTIOBJECTIVE OPTIMIZATION ON A BUDGET: APPLICATION TO MULTIPASS WIRE DRAWING WITH QUANTIFIED UNCERTAINTIES
International Journal for Uncertainty Quantification, Vol.8, 2018, issue 3
Pramod Zagade, Jitesh Panchal, B. P. Gautham, Ilias Bilionis, Piyush Pandita, Amol Joshi
DATA-CONSISTENT SOLUTIONS TO STOCHASTIC INVERSE PROBLEMS USING A PROBABILISTIC MULTI-FIDELITY METHOD BASED ON CONDITIONAL DENSITIES
International Journal for Uncertainty Quantification, Vol.10, 2020, issue 5
L. Bruder, Timothy Wildey, M. W. Gee
MULTI-FIDELITY MODELING OF PROBABILISTIC AERODYNAMIC DATABASES FOR USE IN AEROSPACE ENGINEERING
International Journal for Uncertainty Quantification, Vol.10, 2020, issue 5
Jayant Mukhopadhaya, John F. Quindlen, Brian T. Whitehead, Andrew W. Cary, Juan J. Alonso