On the Use of Upper Trust Bounds in Constrained Bayesian Optimization Infill Criteria

Priem, Rémy; Bartoli, Nathalie; Diouane, Youssef

doi:10.2514/6.2019-2986

Cited by 10 publications

(18 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Concerning the objective function, these information are combined in an acquisition function α (l) f : [17][18][19] coding the trade-off between exploration of the highly uncertain domain that can hide a minimum and exploitation of the minimum of the GP prediction. For the constraints, these information are joined to produce a feasibility criterion α (l) g : 17,20,21] which is generally explicit. The point x (l+1) , solving the constrained maximization trade-off sub-problem:…”

Section: A Bayesian Optimizersmentioning

confidence: 99%

“…WB2S [18]) for highly multimodal objective functions, supplementary feasibility criteria (e.g. the upper trust bound (UTB) [20,21]) for non-linear constraint functions and a multiprocessing ability to speed-up the optimization process. The tool-box [18] is thus named the super efficient global optimization with mixture of experts (SEGOMOE) tool-box.…”

mentioning

confidence: 99%

“…• SEGO [22]: it can be seen as a basic implementation of the super efficient global optimization framework, it was developed for standard constrained optimization problems. • SEGO-UTB [21]: it includes a decreasing upper trust bound in the SEGO framework. In the context of this paper and to encourage the exploration of the feasible domain, we used an exponential decrease for the constraints learning rate, see [21, Fig.…”

mentioning

confidence: 99%

See 2 more Smart Citations

An efficient application of Bayesian optimization to an industrial MDO framework for aircraft design.

Priem¹,

Gagnon

Chittick

et al. 2020

Aiaa Aviation 2020 Forum

Self Cite

View full text Add to dashboard Cite

The multi-level, multidisciplinary and multi-fidelity optimization framework developed at Bombardier Aviation has shown great results to explore efficient and competitive aircraft configurations. This optimization framework has been developed within the Isight software, the latter offers a set of ready-to-use optimizers. Unfortunately, the computational effort required by the Isight optimizers can be prohibitive with respect to the requirements of an industrial context. In this paper, a constrained Bayesian optimization optimizer, namely the super efficient global optimization with mixture of experts, is used to reduce the optimization computational effort. The obtained results showed significant improvements compared to two of the popular Isight optimizers. The capabilities of the tested constrained Bayesian optimization solver are demonstrated on Bombardier research aircraft configuration study cases.

show abstract

Section: A Bayesian Optimizersmentioning

confidence: 99%

mentioning

confidence: 99%

mentioning

confidence: 99%

See 1 more Smart Citation

An efficient application of Bayesian optimization to an industrial MDO framework for aircraft design.

Priem¹,

Gagnon

Chittick

et al. 2020

Aiaa Aviation 2020 Forum

Self Cite

View full text Add to dashboard Cite

show abstract

“…To tackle the issue of badly modeled constraints, including the uncertainties (provided by the GP models) has been shown to be very useful [16][17][18]. For instance, the authors in [16,17] introduced a scalar fixed upper trust bound (UTB) to handle the constraints.…”

Section: Introductionmentioning

confidence: 99%

“…Their proposed approach was designed to allow the exploration of a larger feasible domain (but relaxed) rather than being restricted to a small feasible one. Recently, unlike in [16,17] where only a fixed scalar UTB was investigated, Priem et al [18] used a dynamic adaptive strategy for updating the UTB during the optimization process. The main idea was to include uncertainties on GPs but only during specific stages of the optimization procedure.…”

Section: Introductionmentioning

confidence: 99%

Upper trust bound feasibility criterion for mixed constrained Bayesian optimization with application to aircraft design

Priem

Bartoli

Diouane

et al. 2020

Aerospace Science and Technology

View full text Add to dashboard Cite

Bayesian optimization methods have been successfully applied to black box optimization problems that are expensive to evaluate. In this paper, we adapt the so-called super efficient global optimization algorithm to solve more accurately mixed constrained problems. The proposed approach handles constraints by means of upper trust bound, the latter encourages exploration of the feasible domain by combining the mean prediction and the associated uncertainty function given by the Gaussian processes. On top of that, a refinement procedure, based on a learning rate criterion, is introduced to enhance the exploitation and exploration trade-off. We show the good potential of the approach on a set of numerical experiments. Finally, we present an application to conceptual aircraft configuration upon which we show the superiority of the proposed approach compared to a set of the state-of-the-art black box optimization solvers.

show abstract

A surrogate‐based framework for feasibility‐driven optimization of expensive simulations

Tian,

Ierapetritou

2024

AIChE Journal

View full text Add to dashboard Cite

Surrogate‐based optimization approaches have been widely adopted in industrial problems due to their potential to reduce the number of simulation runs required in the optimization process. The surrogate‐based optimization framework has been extended to feasibility analysis in pharmaceutical manufacturing to characterize the design space. Most surrogate‐based approaches for feasibility analysis are limited to the construction of a regression model for the feasibility function. In this work, we developed a framework with the feasibility problem considered as a classification problem, and additional stages introduced to improve local exploitation and global exploration. We illustrate the efficiency of the proposed framework on three test problems and implement it in a realistic case study describing the production of solid‐based drugs using wet granulation, aimed to reduce the operation cost, improve product quality, and increase process flexibility and robustness.

show abstract

On the Use of Upper Trust Bounds in Constrained Bayesian Optimization Infill Criteria

Cited by 10 publications

References 17 publications

An efficient application of Bayesian optimization to an industrial MDO framework for aircraft design.

An efficient application of Bayesian optimization to an industrial MDO framework for aircraft design.

Upper trust bound feasibility criterion for mixed constrained Bayesian optimization with application to aircraft design

A surrogate‐based framework for feasibility‐driven optimization of expensive simulations

Contact Info

Product

Resources

About