Bayesian optimization using deep Gaussian processes with applications to aerospace system design

DOI10.1007/S11081-020-09517-8zbMATH Open1473.62107arXiv1905.03350OpenAlexW3035278585MaRDI QIDQ2245694FDOQ2245694

Authors: Ali Hebbal, Loïc Brevault, Mathieu Balesdent, El-Ghazali Talbi, Nouredine Melab 

Publication date: 15 November 2021

Published in: Optimization and Engineering (Search for Journal in Brave)

Abstract: Bayesian Optimization using Gaussian Processes is a popular approach to deal with the optimization of expensive black-box functions. However, because of the a priori on the stationarity of the covariance matrix of classic Gaussian Processes, this method may not be adapted for non-stationary functions involved in the optimization problem. To overcome this issue, a new Bayesian Optimization approach is proposed. It is based on Deep Gaussian Processes as surrogate models instead of classic Gaussian Processes. This modeling technique increases the power of representation to capture the non-stationarity by simply considering a functional composition of stationary Gaussian Processes, providing a multiple layer structure. This paper proposes a new algorithm for Global Optimization by coupling Deep Gaussian Processes and Bayesian Optimization. The specificities of this optimization method are discussed and highlighted with academic test cases. The performance of the proposed algorithm is assessed on analytical test cases and an aerospace design optimization problem and compared to the state-of-the-art stationary and non-stationary Bayesian Optimization approaches.

Full work available at URL: https://arxiv.org/abs/1905.03350

Recommendations

• Optimization employing Gaussian process-based surrogates
• Global optimization method using adaptive and parallel ensemble of surrogates for engineering design optimization
• Efficient global optimization for high-dimensional constrained problems by using the Kriging models combined with the partial least squares method
• Selection of a covariance function for a Gaussian random field aimed for modeling global optimization problems
• Survey on Bayesian optimization methodology and applications

zbMATH Keywords

Gaussian processglobal constrained optimizationBayesian optimizationdeep Gaussian processnon-stationary function

Mathematics Subject Classification ID

Computational methods for problems pertaining to statistics (62-08) Nonparametric estimation (62G05) Gaussian processes (60G15)

Cites Work

• Efficient global optimization of expensive black-box functions
• GPflow: a Gaussian process library using TensorFlow
• Bayesian treed Gaussian process models with an application to computer modeling
• Gaussian processes for machine learning.
• Title not available (Why is that?)
• Title not available (Why is that?)
• Efficient global optimization algorithm assisted by multiple surrogate techniques
• A non-stationary covariance-based Kriging method for metamodelling in engineering design
• ORBIT: Optimization by Radial Basis Function Interpolation in Trust-Regions
• On trust region methods for unconstrained minimization without derivatives
• Title not available (Why is that?)
• Warped Gaussian processes and derivative-based sequential designs for functions with heterogeneous variations
• Efficient global optimization for high-dimensional constrained problems by using the Kriging models combined with the partial least squares method

Cited In (9)

• Active Learning for Deep Gaussian Process Surrogates
• Deep Gaussian process for multi-objective Bayesian optimization
• Deep Gaussian Process Emulation using Stochastic Imputation
• Isovolumetric adaptations to space-filling design of experiments
• Optimization employing Gaussian process-based surrogates
• Shape Optimization and Flow Analysis of Supersonic Nozzles Using Deep Learning
• Bayesian optimization for the spanwise oscillation of a gliding flat plate
• Gradient-enhanced deep Gaussian processes for multifidelity modeling
• Sequential design of multi-fidelity computer experiments with effect sparsity

Uses Software

• BOBYQA
• COBYLA2
• EGO
• TensorFlow
• ORBIT
• GPflow
• Adam
• OpenMDAO

This page was built for publication: Bayesian optimization using deep Gaussian processes with applications to aerospace system design

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2245694)