Multi-fidelity Surrogates | springerprofessional.de

Springer Professional

nach oben

2023 | Buch

Kapitel lesen Erstes Kapitel lesen

Multi-fidelity Surrogates

Modeling, Optimization and Applications

verfasst von: Qi Zhou, Min Zhao, Jiexiang Hu, Mengying Ma

Verlag: Springer Nature Singapore

Buchreihe : Engineering Applications of Computational Methods

Enthalten in: Springer Professional "Wirtschaft+Technik" , Springer Professional "Technik" , Springer Professional "Wirtschaft"

Einloggen, um Zugang zu erhalten

Über dieses Buch

This book investigates two types of static multi-fidelity surrogates modeling approaches, sequential multi-fidelity surrogates modeling approaches, the multi-fidelity surrogates-assisted efficient global optimization, reliability analysis, robust design optimization, and evolutionary optimization. Multi-fidelity surrogates have attracted a significant amount of attention in simulation-based design and optimization in recent years. Some real-life engineering design problems, such as prediction of angular distortion in the laser welding, optimization design of micro-aerial vehicle fuselage, and optimization design of metamaterial vibration isolator, are also provided to illustrate the ability and merits of multi-fidelity surrogates in support of engineering design. Specifically, lots of illustrative examples are adopted throughout the book to help explain the approaches in a more “hands-on” manner. This book is a useful reference for postgraduates and researchers of mechanical engineering, as well as engineers of enterprises in related fields.

Anzeige

Inhaltsverzeichnis

Frontmatter

Chapter 1. Introduction

Abstract

Physics-based simulation models from different disciplines are becoming indispensable in modern product design. In the preliminary design phase, these simulation models can help predict the performance of products to expedite the design space exploration and search for the optimal design.

Qi Zhou, Min Zhao, Jiexiang Hu, Mengying Ma

Chapter 2. Hierarchical Multi-fidelity Surrogate Modeling

Abstract

Most of the existing multi-fidelity (MF) surrogates assume that high-fidelity (HF) models are generally more accurate than low-fidelity (LF) models but LF models are less expensive than HF models. In other words, the fidelity levels between the HF and LF models can be clearly identified, which is the core concept of hierarchical multi-fidelity surrogate modeling. The motivation of MF surrogate modeling is that a large number of inexpensive LF sampling points can be used to decrease the computational cost, while a limited number of expensive HF sampling points can be used to ensure the prediction accuracy of the surrogate model.

Qi Zhou, Min Zhao, Jiexiang Hu, Mengying Ma

Chapter 3. Nonhierarchical Multi-fidelity Surrogate Modeling

Abstract

The methods of constructing multi-fidelity (MF) surrogates can be divided into two categories: hierarchical and nonhierarchical methods. Several hierarchical methods have been described in Chap. 2.

Qi Zhou, Min Zhao, Jiexiang Hu, Mengying Ma

Chapter 4. Sequential Multi-fidelity Surrogate Modeling

Abstract

Under a limited computational budget, the quality of a multi-fidelity (MF) surrogate depends on the distributions of the sample points and sample size ratio between the low-fidelity (LF) and high-fidelity (HF) simulations, i.e., the allocation of the computational budget.

Qi Zhou, Min Zhao, Jiexiang Hu, Mengying Ma

Chapter 5. Multi-fidelity Surrogate Assisted Efficient Global Optimization

Abstract

As reported in the previous literature, multi-fidelity (MF) surrogate assisted design optimization techniques can be classified into two types: offline and online techniques. In the offline technique, a prespecified number of sample points is used to build an MF model to replace the simulation analysis in the engineering optimization process.

Qi Zhou, Min Zhao, Jiexiang Hu, Mengying Ma

Chapter 6. Multi-fidelity Surrogate Assisted Reliability Design Optimization

Abstract

The structural reliability problem has received increasing attention with the increasing complexity of engineering structures. Failure probability is the main issue considered in structural reliability analyses. Multiple integrals must be evaluated to analytically obtain the failure probability.

Qi Zhou, Min Zhao, Jiexiang Hu, Mengying Ma

Chapter 7. Multi-fidelity Surrogate Assisted Robust Design Optimization

Abstract

Engineering product design optimization inevitably involves uncertainties, which may degrade the objective performance or render the optimal solution infeasible. To alleviate the sensitivity of the performance of engineering products to uncertainty factors, the influence of uncertainties must be comprehensively studied in the design stage. Robust optimization (RO) methods aim to ensure the key performances of engineering products and alleviate their sensitivity to uncertainty factors.

Qi Zhou, Min Zhao, Jiexiang Hu, Mengying Ma

Chapter 8. Multi-fidelity Surrogate Assisted Evolutional Optimization

Abstract

Evolution algorithms, such as multi-objective genetic algorithms (MOGAs), require a large number of function evaluations to converge to global optima or near-optimal solutions (Sun et al. in IEEE Trans Cybern 43:685–698, 2013; Cheng et al. in IEEE Trans Evol Comput 19:838–856, 2015). This aspect limits, to a certain extent, their capability of solving real-world engineering design problems, which typically involve computationally expensive simulation models, i.e., HF models. The following three strategies are commonly used to increase the algorithm efficiency: The first strategy pertains to fitness inheritance.

Qi Zhou, Min Zhao, Jiexiang Hu, Mengying Ma

Chapter 9. Engineering Applications

Abstract

Multi-fidelity (MF) surrogates can balance the prediction accuracy and computational cost by augmenting a few expensive high-fidelity (HF) samples with many inexpensive low-fidelity (LF) data. Consequently, MF models have been widely applied in engineering design and optimization.

Qi Zhou, Min Zhao, Jiexiang Hu, Mengying Ma

Chapter 10. Concluding Remarks

Abstract

Chapter 1 introduces the concept of multi-fidelity surrogates, providing the readers with a general understanding of what such surrogates are and their expected use cases. Subsequently, the state-of-the-art techniques for multi-fidelity surrogate modeling and optimization, and their applications in mechanical design and optimization are summarized.

Qi Zhou, Min Zhao, Jiexiang Hu, Mengying Ma

Titel: Multi-fidelity Surrogates
verfasst von: Qi Zhou
Min Zhao
Jiexiang Hu
Mengying Ma
Copyright-Jahr: 2023
Verlag: Springer Nature Singapore
Electronic ISBN: 978-981-19-7210-2
Print ISBN: 978-981-19-7209-6
DOI: https://doi.org/10.1007/978-981-19-7210-2

Premium Partner

Marktübersichten

Die im Laufe eines Jahres in der „adhäsion“ veröffentlichten Marktübersichten helfen Anwendern verschiedenster Branchen, sich einen gezielten Überblick über Lieferantenangebote zu verschaffen.

Zur Marktübersicht