Machine learning-based surrogate modeling for data-driven optimization: a comparison of subset selection for regression techniques

Kim, Sun Hye; Boukouvala, Fani

doi:10.1007/s11590-019-01428-7

Cited by 68 publications

(47 citation statements)

References 44 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Future extensions of GB-EPI could include adaptive meshing of the centroidal Voronoi tessellations 51 to increase the number of niches in the most suitable regions of feature space, surrogate modelling techniques 52,53 to reduce the number of necessary tness function evaluations, or crossovers based on intermolecular correlations. 54 In addition, deep learning models could be trained to predict which mutations are most benecially applied to which molecules.…”

Section: Discussionmentioning

confidence: 99%

Illuminating elite patches of chemical space

Verhellen

Abeele

2020

Chem. Sci.

View full text Add to dashboard Cite

show abstract

Section: Discussionmentioning

confidence: 99%

Illuminating elite patches of chemical space

Verhellen

Abeele

2020

Chem. Sci.

View full text Add to dashboard Cite

show abstract

“…Currently, multiple types of surrogate models can be utilized such as linear regression, kriging, artificial neural network (ANN), radial basis function, and so forth. Among them, some surrogate models (e.g., random forest) cannot provide available derivative information while the derivatives of many other surrogate models are symbolically available such as linear regression, ANN with tansig kernel function, and so forth 39 . Here, a surrogate model with available derivative information is preferred because solving a discrete‐continuous optimization problem with no derivative information is very challenging.…”

Section: Systematic Procedures For Optimization Problem Formulationmentioning

confidence: 99%

“…The hyperparameters of the surrogate model structure should be carefully tuned. The heuristics and experience reported in the literature can be consulted 39,40 . Afterward, model accuracy needs to be validated.…”

Section: Systematic Procedures For Optimization Problem Formulationmentioning

confidence: 99%

Optimization‐based cosmetic formulation: Integration of mechanistic model, surrogate model, and heuristics

Zhang

Zhou

2020

AIChE Journal

View full text Add to dashboard Cite

Multiple functional and hard‐to‐quantify sensorial product attributes that can be satisfied by a large number of cosmetic ingredients are required in the design of cosmetics. To overcome this problem, a new optimization‐based approach for expediting cosmetic formulation is presented. It exploits the use of a hierarchy of models in an iterative manner to refine the search for creating the highest‐quality cosmetic product. First, a systematic procedure is proposed for optimization problem formulation, where the cosmetic formulation problem is defined, design variables are specified, and a set of models for sensorial perception and desired product properties are identified. Then, a solution strategy that involves iterative model adoption and two numerical techniques (i.e., generalized disjunctive programming reformulation and model substitution) is applied to improve the efficiency of solving the optimization problem and to find better solutions. The applicability of the proposed procedure and solution strategy is illustrated with a perfume formulation example.

show abstract

“…One method to propagate uncertainty is using models or surrogate models. A surrogate model is a cheap-to-run approximation of the actual model (Kim and Boukouvala, 2019). Among the surrogate models, Polynomial Chaos Expansion (PCE) has gained attention especially after the work of Ghanem and Spanos (2003) and Xiu and Karniadakis (2003).…”

Section: Polynomial Chaos Expansionmentioning

confidence: 99%

Surrogate models for unsteady aerodynamics using non-intrusive Polynomial Chaos Expansions

Haghi

Crawford

2020

Preprint

View full text Add to dashboard Cite

Abstract. In common industrial practice based on IEC standards, wind turbine simulations are computed in the time domain for each mean wind speed bin using six unsteady wind seeds. Different software such as FAST, Balded or HAWC2 can be used to this purpose, to capture the unsteadiness and uncertainties of the wind in the simulations. The statistics of these simulations are extracted and used to calculate fatigue and extreme loads on the wind turbine components. Having only six seeds does not guarantee an accurate estimation of the overall statistics. One solution might be running more seeds; however, this will increase the computation cost. Moreover, to move beyond Blade Element Momentum based tools toward vortex/potential flow formulations, a reduction in the computational cost associated with the unsteady flow and uncertainty handling is required. This study illustrates the stationary character of the unsteady aerodynamic statistics based on the standard turbulence models. Afterward, we propose a non-intrusive Polynomial Chaos Expansion to build a surrogate model of the loads' statistics at each time step, to estimate the statistics more accurately and efficiently.

show abstract

Machine learning-based surrogate modeling for data-driven optimization: a comparison of subset selection for regression techniques

Cited by 68 publications

References 44 publications

Illuminating elite patches of chemical space

Illuminating elite patches of chemical space

Optimization‐based cosmetic formulation: Integration of mechanistic model, surrogate model, and heuristics

Surrogate models for unsteady aerodynamics using non-intrusive Polynomial Chaos Expansions

Contact Info

Product

Resources

About