Distributed Gauss-Newton optimization method for history matching problems with multiple best matches

Gao, Guohua; Vink, Jeroen C.; Chen, Chaohui; Khamra, Yaakoub El; Tarrahi, Mohammadali

doi:10.1007/s10596-017-9657-9

Cited by 34 publications

(6 citation statements)

References 59 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The under-performance of plain, ensemble-based algorithms in handling multi-modal variables is also discussed in the literature, see, for example, Elsheikh et al (2013); Gao et al (2017Gao et al ( , 2018; of kernel parameters that are associated with the center points (note that different clusters of training data share the same set of center points). The corrected predictions are then taken as biased outputs plus certain residual terms, whereas the latter are calculated as the weighted averages of the residuals predicted using the kernel parameters in each cluster.…”

Section: Results With Respect To Multi-modal Inputsmentioning

confidence: 99%

Ensemble-based kernel learning for a class of data assimilation problems with imperfect forward simulators

Luo

2019

PLoS ONE

View full text Add to dashboard Cite

Simulator imperfection, often known as model error, is ubiquitous in practical data assimilation problems. Despite the enormous efforts dedicated to addressing this problem, properly handling simulator imperfection in data assimilation remains to be a challenging task. In this work, we propose an approach to dealing with simulator imperfection from a point of view of functional approximation that can be implemented through a certain machine learning method, such as kernel-based learning adopted in the current work. To this end, we start from considering a class of supervised learning problems, and then identify similarities between supervised learning and variational data assimilation. These similarities found the basis for us to develop an ensemble-based learning framework to tackle supervised learning problems, while achieving various advantages of ensemble-based methods over the variational ones. After establishing the ensemble-based learning framework, we proceed to investigate the integration of ensemble-based learning into an ensemble-based data assimilation framework to handle simulator imperfection. In the course of our investigations, we also develop a strategy to tackle the issue of multi-modality in supervised-learning problems, and transfer this strategy to data assimilation problems to help improve assimilation performance. For demonstration, we apply the ensemble-based learning framework and the integrated, ensemble-based data assimilation framework to a supervised learning problem and a data assimilation problem with an imperfect forward simulator, respectively. The experiment results indicate that both frameworks achieve good performance in relevant case studies, and that functional approximation through machine learning may serve as a viable way to account for simulator imperfection in data assimilation problems.

show abstract

Section: Results With Respect To Multi-modal Inputsmentioning

confidence: 99%

Ensemble-based kernel learning for a class of data assimilation problems with imperfect forward simulators

Luo

2019

PLoS ONE

View full text Add to dashboard Cite

show abstract

“…While EXPLO2 is parallelized (at marginal cost to performance), to the best of our knowledge no competing technique is except for * -CMA-ES. Indeed, as (Gao et al, 2017) points out, most surrogate-based derivative-free optimizersincluding NEWUOA-require sequential function evaluations. Though parallel algorithms exist (Haftka et al, 2016;Rehbach et al, 2018;Xia & Shoemaker, 2020), these are relatively few in number outside the context of Bayesian optimization (which is unsuitable for high-dimensional problems), and parallel techniques appropriate for highdimensional problems have been considered in our design and/or benchmarking.…”

Section: Parallel Performancementioning

confidence: 99%

Parallel black-box optimization of expensive high-dimensional multimodal functions via magnitude

Huntsman¹

2022

Preprint

View full text Add to dashboard Cite

Building on the recently developed theory of magnitude, we introduce the optimization algorithm EXPLO2 and carefully benchmark it. EXPLO2 advances the state of the art for optimizing highdimensional (D ⪆ 40) multimodal functions that are expensive to compute and for which derivatives are not available, such as arise in hyperparameter optimization or via simulations. Example 1. Let {x j } 3 j=1 ⊂ R 2 have pairwise distances d jk ∶= d(x j , x k ) given by d 12 = d 13 = 1 = d 21 = d 31 and d 23 = δ = d 32 with δ < 2. It turns out that w 1 = e (δ+2)t − 2e (δ+1)t + e 2t e (δ+2)t − 2e δt + e 2t ;

show abstract

“…While EXPLO2 is parallelized (at marginal cost to performance), to the best of our knowledge no competing technique is except for *-CMA-ES. Indeed, as [6] points out, most surrogate-based derivativefree optimizers-including NEWUOA-require sequential function evaluations. Though parallel algorithms exist [7,18,25], these are relatively few in number outside the context of Bayesian optimization (which is unsuitable for high-dimensional problems), and parallel techniques appropriate for high-dimensional problems have been considered in our design and/or benchmarking.…”

Section: Parallel Performancementioning

confidence: 99%

Benchmarking an algorithm for expensive high-dimensional objectives on the bbob and bbob-largescale testbeds

Hoffman

Huntsman

2022

Proceedings of the Genetic and Evolutionary Computation Conference Companion

View full text Add to dashboard Cite

We report benchmarks for a recently developed algorithm on the bbob and bbob-largescale benchmarking testbeds in COCO. This algorithm is designed for expensive high-dimensional multimodal objectives (such as arise in hyperparameter optimization or via simulations), and this regime introduces challenges for benchmarking. In particular, while the COCO experimental procedure yields evidence that this algorithm improves on the state of the art, the COCO framework also exhibits shortcomings for the very low evaluation budgets involved here. Consequently, we also report the results of ad hoc experiments that demonstrate a more obvious if less rigorous advantage for the algorithm over its nearest competition. CCS CONCEPTS• Computing methodologies → Continuous space search.

show abstract

Distributed Gauss-Newton optimization method for history matching problems with multiple best matches

Cited by 34 publications

References 59 publications

Ensemble-based kernel learning for a class of data assimilation problems with imperfect forward simulators

Ensemble-based kernel learning for a class of data assimilation problems with imperfect forward simulators

Parallel black-box optimization of expensive high-dimensional multimodal functions via magnitude

Benchmarking an algorithm for expensive high-dimensional objectives on the bbob and bbob-largescale testbeds

Contact Info

Product

Resources

About