A multi-fidelity surrogate model based on support vector regression

Shi, Maolin; Lv, Liye; Sun, Wei; Song, Xueguan

doi:10.1007/s00158-020-02522-6

Cited by 60 publications

(8 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Yi et al (2020b) presented a multi-fidelity RBF surrogate-based optimization framework, where the LF and HF surrogate models are sequentially exploited. Apart from RBF, support vector regression (SVR) model was also employed to build VF surrogate models (Shi et al 2020a). To accelerate the VFSBO, researchers begins to develop the parallel VF optimization methods, for instance, He et al (2021) extended the VF-EI approach to a parallel process in consideration of simulation failures.…”

Section: Related Workmentioning

confidence: 99%

An enhanced variable-fidelity optimization approach for constrained optimization problems and its parallelization

Lin

2022

Struct Multidisc Optim

View full text Add to dashboard Cite

In this paper, a variable-fidelity constrained lower confidence bound (VF-CLCB) criterion is presented for computationally expensive constrained optimization problems (COPs) with two levels of fidelity. In VF-CLCB, the hierarchical Kriging model is adopted to model the objective and inequality constraints. Two infill sampling functions are developed based on the objective and the constraints, respectively, and an adaptive selection strategy is set to select the elite sample points. Moreover, based on the VF-CLCB criterion, a parallel optimization method noted as PVF-CLCB is subsequently developed to accelerate the optimization process. In PVF-CLCB, a VF influence function is defined to approximately evaluate the estimation error of the hierarchical Kriging models, based on which multiple promising points can be determined at each iteration. In addition, an allocation strategy is proposed to distribute the computation resources between the objective- and constraint-oriented functions properly. Lastly, the proposed VF-CLCB and PVF-CLCB approaches are compared with the alternative methods on 12 benchmark numerical cases, and their significant superiority in solving computationally expensive COPs is verified. Furthermore, the proposed methods are employed to optimize the global stability of the stiffened cylindrical shell, and the optimum structure is yielded.

show abstract

Section: Related Workmentioning

confidence: 99%

An enhanced variable-fidelity optimization approach for constrained optimization problems and its parallelization

Lin

2022

Struct Multidisc Optim

View full text Add to dashboard Cite

show abstract

“…Besides the Gaussian processes implemented in this work, other surrogate model approaches are often examined and enhanced. Popular models are radial-basis functions (Song et al 2019) and support vector regression (Shi et al 2020), which are relatively performant and precise. There are also studies evaluating neural networks (Springenberg et al 2016;Alam et al 2004) or even splines (Turner and Crawford 2008).…”

Section: Gaussian Processesmentioning

confidence: 99%

Spline-based shape optimization of large-scale composite leaf spring models using Bayesian strategies with multiple constraints

Winter

Fiebig²,

Franke³

et al. 2022

Struct Multidisc Optim

View full text Add to dashboard Cite

The presented paper describes a shape optimization workflow using Bayesian strategies. It is applied to a novel automotive axle system consisting of leaf springs made from glass fiber reinforced plastics (GFRP). Besides the primary objectives of cost and mass reduction, the assembly has to meet multiple technical constraints with respect to various loading conditions. The related large-scale finite element model is fully parameterized by splines, hence the general shape of the guide curve as well as the spring’s height, width and material properties can be altered by the corresponding workflow. For this purpose, a novel method is developed to automatically generate high-quality meshes depending on the geometry of the respective springs. The size and complexity of the model demands the implementation of efficient optimization techniques with a preferably small number of required response function evaluations. Therefore, an existing optimization framework is extended by state-of-the-art Bayesian methods, including different kernel combinations and multiple acquisition function approaches, which are then tested, evaluated and compared. To properly address the use of GFRP as spring material in the objective function, an appropriate cost model is derived. Emerging challenges, such as conflicting targets regarding direct material costs and potential lightweight measures, are considered and investigated. The intermediate steps of the developed optimization procedure are tested on various sample functions and simplified models. The entire workflow is finally applied to the complete model and evaluated. Concluding, ideas and possibilities in improving the optimization process, such as the use of models with varying complexity, are discussed.

show abstract

“…Zhang et al (2018) used the LF model to evaluate the trend of the HF model and estimated the coefficients of the LF model and discrepancy function using linear regression. Shi et al (2020) proposed an MFS model based on support vector regression, in which the correlation between LF and HF models is described in a mapped high-dimensional space. Rumpfkeil et al (2019) proposed an MFS model based on sparse polynomial chaos expansion.…”

Section: Introductionmentioning

confidence: 99%

A multi-fidelity surrogate model based on extreme support vector regression: fusing different fidelity data for engineering design

Shi

2023

Self Cite

View full text Add to dashboard Cite

PurposeExtreme support vector regression (ESVR) has been widely used in the design, analysis and optimization of engineering systems of its fast training speed and good computational ability. However, the ESVR model is only able to utilize one-fidelity information of engineering system. To solve this issue, this paper extends extreme support vector regression (ESVR) to a multi-fidelity surrogate (MFS) model which can make use of a few expensive but higher-fidelity (HF) samples and a lot of inaccurate but cheap low-fidelity (LF) samples, named ESVR-MFS.Design/methodology/approachIn the ESVR-MFS model, a kernel matrix is designed to evaluate the relationship between the HF and LF samples. The root mean square error of HF samples is used as the training error metric, and the optimal hyper-parameters of the kernel matrix are obtained through a heuristic algorithm.FindingsA number of numerical problems and three engineering problems are used to compare the ESVR-MFS model with the single-fidelity ESVR model and two benchmark MFS models. The results show that the ESVR-MFS model exhibits competitive performance in both numerical cases and practical cases tested in this work.Practical implicationsThe proposed approach exhibits great capability for practical multi-fidelity engineering design problems.Originality/valueA MFS model is proposed based on ESVR, which can make full use of the advantages of both HF data and LF data to achieve optimal results at same or lower cost.

show abstract

A multi-fidelity surrogate model based on support vector regression

Cited by 60 publications

References 42 publications

An enhanced variable-fidelity optimization approach for constrained optimization problems and its parallelization

An enhanced variable-fidelity optimization approach for constrained optimization problems and its parallelization

Spline-based shape optimization of large-scale composite leaf spring models using Bayesian strategies with multiple constraints

A multi-fidelity surrogate model based on extreme support vector regression: fusing different fidelity data for engineering design

Contact Info

Product

Resources

About