A multi-fidelity Bayesian optimization approach based on the expected further improvement

Sampling efficiency is important for simulation-based design optimization. While Bayesian optimization (BO) has been successfully applied in engineering problems, the cost associated with large-scale simulations has not been fully addressed. Extending the standard BO approaches to multi-fidelity optimization can utilize the information of low-fidelity models to further reduce the optimization cost. In this work, a multi-fidelity Bayesian optimization approach is proposed, in which hierarchical Kriging is used for constructing the multi-fidelity metamodel. The proposed approach quantifies the effect of HF and LF samples in multi-fidelity optimization based on a new concept of expected further improvement. A novel acquisition function is proposed to determine both the location and fidelity level of the next sample simultaneously, with the consideration of balance between the value of information provided by the new sample and the associated sampling cost. The proposed approach is compared with some state-of-the-art methods for multi-fidelity global optimization with numerical examples and an engineering case. The results show that the proposed approach can obtain global optimal solutions with reduced computational costs.

[1]  Douglas Allaire,et al.  Multi-information source constrained Bayesian optimization , 2018, Structural and Multidisciplinary Optimization.

[2]  Stefano Ermon,et al.  Sparse Gaussian Processes for Bayesian Optimization , 2016, UAI.

[3]  Raphael T. Haftka,et al.  Surrogate-based Analysis and Optimization , 2005 .

[4]  Carlos A. Coello Coello,et al.  Use of a self-adaptive penalty approach for engineering optimization problems , 2000 .

[5]  Yu Zhang,et al.  Variable-fidelity expected improvement method for efficient global optimization of expensive functions , 2018 .

[6]  刘俊,et al.  Efficient Variable-fidelity Optimization Applied to Benchmark Transonic Airfoil Design , 2015 .

[7]  Hui Zhou,et al.  An adaptive global variable fidelity metamodeling strategy using a support vector regression based scaling function , 2015, Simul. Model. Pract. Theory.

[8]  S. Koziel,et al.  A Space-Mapping Framework for Engineering Optimization—Theory and Implementation , 2006, IEEE Transactions on Microwave Theory and Techniques.

[9]  G. G. Wang,et al.  Adaptive Response Surface Method Using Inherited Latin Hypercube Design Points , 2003 .

[10]  Liang Gao,et al.  An enhanced RBF-HDMR integrated with an adaptive sampling method for approximating high dimensional problems in engineering design , 2016 .

[11]  Anh Tran,et al.  sMF-BO-2CoGP: A Sequential Multi-Fidelity Constrained Bayesian Optimization Framework for Design Applications , 2020, J. Comput. Inf. Sci. Eng..

[12]  Tom Dhaene,et al.  Performance study of multi-fidelity gradient enhanced kriging , 2015 .

[13]  Pramudita Satria Palar,et al.  Multi-Fidelity Uncertainty Analysis in CFD Using Hierarchical Kriging , 2017 .

[14]  Donald R. Jones,et al.  Efficient Global Optimization of Expensive Black-Box Functions , 1998, J. Glob. Optim..

[15]  Yan Wang,et al.  A sequential multi-fidelity metamodeling approach for data regression , 2017, Knowl. Based Syst..

[16]  Leshi Shu,et al.  An online variable-fidelity optimization approach for multi-objective design optimization , 2019, Structural and Multidisciplinary Optimization.

[17]  Wei Chen,et al.  A New Variable-Fidelity Optimization Framework Based on Model Fusion and Objective-Oriented Sequential Sampling , 2007, DAC 2007.

[18]  R. A. Miller,et al.  Sequential kriging optimization using multiple-fidelity evaluations , 2006 .

[19]  A. O'Hagan,et al.  Predicting the output from a complex computer code when fast approximations are available , 2000 .

[20]  Seyede Fatemeh Ghoreishi,et al.  Efficient Use of Multiple Information Sources in Material Design , 2019, Acta Materialia.

[21]  G. Gary Wang,et al.  Review of Metamodeling Techniques in Support of Engineering Design Optimization , 2007 .

[22]  Shawn E. Gano,et al.  Hybrid Variable Fidelity Optimization by Using a Kriging-Based Scaling Function , 2005 .

[23]  Ankit Srivastava,et al.  Multi-Information Source Fusion and Optimization to Realize ICME: Application to Dual-Phase Materials , 2018, Journal of Mechanical Design.

[24]  Chao-Chao Wang,et al.  An active learning metamodeling approach by sequentially exploiting difference information from variable-fidelity models , 2016, Adv. Eng. Informatics.

[25]  Kwanjung Yee,et al.  High-to-Low Initial Sample Ratio of Hierarchical Kriging for Film Hole Array Optimization , 2018 .

[26]  John W. Bandler,et al.  Space mapping technique for electromagnetic optimization , 1994 .

[27]  Donald R. Jones,et al.  Global versus local search in constrained optimization of computer models , 1998 .

[28]  Jeong‐Soo Park.  Optimal Latin-hypercube designs for computer experiments , 1994 .

[29]  John W. Bandler,et al.  An Introduction to the Space Mapping Technique , 2001 .

[30]  Yixin Liu,et al.  Sequential optimization using multi-level cokriging and extended expected improvement criterion , 2018 .

[31]  Pierre-Alain Boucard,et al.  Variable-fidelity modeling of structural analysis of assemblies , 2016, J. Glob. Optim..

[32]  Qi Zhou,et al.  Novel Approach for Selecting Low-Fidelity Scale Factor in Multifidelity Metamodeling , 2019 .

[33]  Yan Wang,et al.  pBO-2GP-3B: A batch parallel known/unknown constrained Bayesian optimization with feasibility classification and its applications in computational fluid dynamics , 2019, Computer Methods in Applied Mechanics and Engineering.

[34]  Minh Tran,et al.  Constrained mixed-integer Gaussian mixture Bayesian optimization and its applications in designing fractal and auxetic metamaterials , 2019, Structural and Multidisciplinary Optimization.

[35]  Xudong Yang,et al.  Gradient-Enhanced Hierarchical Kriging Model for Aerodynamic Design Optimization , 2017 .

[36]  Donald R. Jones,et al.  A Taxonomy of Global Optimization Methods Based on Response Surfaces , 2001, J. Glob. Optim..

[37]  Stefan Görtz,et al.  Hierarchical Kriging Model for Variable-Fidelity Surrogate Modeling , 2012 .

[38]  Yahui Zhang,et al.  Metamodel-based design optimization employing a novel sequential sampling strategy , 2017 .

[39]  Raphael T. Haftka,et al.  Sensitivity-based scaling for approximating. Structural response , 1993 .