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 estimation problems arising in detailed combustion kinetics.Comment: Preprint 53 pages, 17 figures (54 small figures). v1 submitted to the Journal of Computational Physics on August 4, 2011; v2 submitted on August 12, 2012. v2 changes: (a) addition of Appendix B and Figure 17 to address the bias in the expected utility estimator; (b) minor language edits; v3 submitted on November 30, 2012. v3 changes: minor edit
1 January 2014 in International Journal for Uncertainty Quantification
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.
Questions of ‘how best to acquire data’ are essential to modelling and prediction in the natural and social sciences, engineering applications, and beyond. Optimal experimental design (OED) formalizes these questions and creates computational methods to answer them. This article presents a systematic survey of modern OED, from its foundations in classical design theory to current research involving OED for complex models. We begin by reviewing criteria used to formulate an OED problem and thus to encode the goal of performing an experiment. We emphasize the flexibility of the Bayesian and decision-theoretic approach, which encompasses information-based criteria that are well-suited to nonlinear and non-Gaussian statistical models. We then discuss methods for estimating or bounding the values of these design criteria; this endeavour can be quite challenging due to strong nonlinearities, high parameter dimension, large per-sample costs, or settings where the model is implicit. A complementary set of computational issues involves optimization methods used to find a design; we discuss such methods in the discrete (combinatorial) setting of observation selection and in settings where an exact design can be continuously parametrized. Finally we present emerging methods for sequential OED that build non-myopic design policies, rather than explicit designs; these methods naturally adapt to the outcomes of past experiments in proposing new experiments, while seeking coordination among all experiments to be performed. Throughout, we highlight important open questions and challenges.
Digital Twins (DTs) are receiving considerable attention from multiple disciplines. Much of the literature at this time is dedicated to the conceptualization of digital twins, and associated enabling technologies and challenges. In this paper, we consider these propositions for the specific application of nuclear power. Our review finds that the current DT concepts are amenable to nuclear power systems, but benefit from some modifications and enhancements. Further, some areas of the existing modeling and simulation infrastructure around nuclear power systems are adaptable to DT development, while more recent efforts in advanced modeling and simulation are less suitable at this time. For nuclear power applications, DT development should rely first on mechanistic model-based methods to leverage the extensive experience and understanding of these systems. Model-free techniques can then be adopted to selectively, and correctively, augment limitations in the model-based approaches. Challenges to the realization of a DT are also discussed, with some being unique to nuclear engineering, however most are broader. A challenging aspect we discuss in detail for DTs is the incorporation of uncertainty quantification (UQ). Forward UQ enables the propagation of uncertainty from the digital representations to predict behavior of the physical asset. Similarly, inverse UQ allows for the incorporation of data from new measurements obtained from the physical asset back into the DT. Optimization under uncertainty facilitates decision support through the formal methods of optimal experimental design and design optimization that maximize information gain, or performance, of the physical asset in an uncertain environment.
1 November 2019 in Computer Methods in Applied Mechanics and Engineering