Simulation-based optimal Bayesian experimental design for nonlinear systems
The optimal selection of experimental conditions is essential to maximizing the value of data for inference and prediction, particularly in situations where experiments are time-consuming and expensive to conduct. We propose a general mathematical framework and an algorithmic approach for optimal ex...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | en_US |
| Published: |
Elsevier
2015
|
| Online Access: | http://hdl.handle.net/1721.1/99467 https://orcid.org/0000-0001-6544-2764 https://orcid.org/0000-0001-8242-3290 |
| _version_ | 1826209224257961984 |
|---|---|
| author | Huan, Xun Marzouk, Youssef M. |
| author2 | Massachusetts Institute of Technology. Department of Aeronautics and Astronautics |
| author_facet | Massachusetts Institute of Technology. Department of Aeronautics and Astronautics Huan, Xun Marzouk, Youssef M. |
| author_sort | Huan, Xun |
| collection | MIT |
| description | The optimal selection of experimental conditions is essential to maximizing the value of data for inference and prediction, particularly in situations where experiments are time-consuming and expensive to conduct. We propose a general mathematical framework and an algorithmic approach for optimal experimental design with nonlinear simulation-based models; in particular, we focus on finding sets of experiments that provide the most information about targeted sets of parameters.
Our framework employs a Bayesian statistical setting, which provides a foundation for inference from noisy, indirect, and incomplete data, and a natural mechanism for incorporating heterogeneous sources of information. An objective function is constructed from information theoretic measures, reflecting expected information gain from proposed combinations of experiments. Polynomial chaos approximations and a two-stage Monte Carlo sampling method are used to evaluate the expected information gain. Stochastic approximation algorithms are then used to make optimization feasible in computationally intensive and high-dimensional settings. These algorithms are demonstrated on model problems and on nonlinear parameter inference problems arising in detailed combustion kinetics. |
| first_indexed | 2024-09-23T14:19:09Z |
| format | Article |
| id | mit-1721.1/99467 |
| institution | Massachusetts Institute of Technology |
| language | en_US |
| last_indexed | 2024-09-23T14:19:09Z |
| publishDate | 2015 |
| publisher | Elsevier |
| record_format | dspace |
| spelling | mit-1721.1/994672022-09-29T08:40:22Z Simulation-based optimal Bayesian experimental design for nonlinear systems Huan, Xun Marzouk, Youssef M. Massachusetts Institute of Technology. Department of Aeronautics and Astronautics Huan, Xun Marzouk, Youssef M. The optimal selection of experimental conditions is essential to maximizing the value of data for inference and prediction, particularly in situations where experiments are time-consuming and expensive to conduct. We propose a general mathematical framework and an algorithmic approach for optimal experimental design with nonlinear simulation-based models; in particular, we focus on finding sets of experiments that provide the most information about targeted sets of parameters. Our framework employs a Bayesian statistical setting, which provides a foundation for inference from noisy, indirect, and incomplete data, and a natural mechanism for incorporating heterogeneous sources of information. An objective function is constructed from information theoretic measures, reflecting expected information gain from proposed combinations of experiments. Polynomial chaos approximations and a two-stage Monte Carlo sampling method are used to evaluate the expected information gain. Stochastic approximation algorithms are then used to make optimization feasible in computationally intensive and high-dimensional settings. These algorithms are demonstrated on model problems and on nonlinear parameter inference problems arising in detailed combustion kinetics. King Abdullah University of Science and Technology (Global Research Partnership) United States. Dept. of Energy. Office of Advanced Scientific Computing Research (Grant DE-SC0003908) 2015-10-27T14:14:35Z 2015-10-27T14:14:35Z 2012-09 2012-08 Article http://purl.org/eprint/type/JournalArticle 00219991 1090-2716 http://hdl.handle.net/1721.1/99467 Huan, Xun, and Youssef M. Marzouk. “Simulation-Based Optimal Bayesian Experimental Design for Nonlinear Systems.” Journal of Computational Physics 232, no. 1 (January 2013): 288–317. https://orcid.org/0000-0001-6544-2764 https://orcid.org/0000-0001-8242-3290 en_US http://dx.doi.org/10.1016/j.jcp.2012.08.013 Journal of Computational Physics Creative Commons Attribution-Noncommercial-NoDerivatives http://creativecommons.org/licenses/by-nc-nd/4.0/ application/pdf Elsevier Arxiv |
| spellingShingle | Huan, Xun Marzouk, Youssef M. Simulation-based optimal Bayesian experimental design for nonlinear systems |
| title | Simulation-based optimal Bayesian experimental design for nonlinear systems |
| title_full | Simulation-based optimal Bayesian experimental design for nonlinear systems |
| title_fullStr | Simulation-based optimal Bayesian experimental design for nonlinear systems |
| title_full_unstemmed | Simulation-based optimal Bayesian experimental design for nonlinear systems |
| title_short | Simulation-based optimal Bayesian experimental design for nonlinear systems |
| title_sort | simulation based optimal bayesian experimental design for nonlinear systems |
| url | http://hdl.handle.net/1721.1/99467 https://orcid.org/0000-0001-6544-2764 https://orcid.org/0000-0001-8242-3290 |
| work_keys_str_mv | AT huanxun simulationbasedoptimalbayesianexperimentaldesignfornonlinearsystems AT marzoukyoussefm simulationbasedoptimalbayesianexperimentaldesignfornonlinearsystems |