Multidisciplinary design optimization in computational mechanics [[electronic resource] /] / edited by Piotr Breitkopf, Rajan Filomeno Coelho

• Record Nr.: UNINA9910141601803321
• Pubbl/distr/stampa: London, : ISTE; Hoboken, N.J., : Wiley, 2010
• ISBN: 1-118-60015-0; 1-118-60002-9
• Descrizione fisica: 1 online resource (573 p.)
• Collana: ISTE
• Altri autori (Persone): BreitkopfPiotr; CoelhoRajan Filomeno
• Disciplina: 620.100285
• Soggetti: Engineering design; Engineering mathematics
• Lingua di pubblicazione: Inglese
• Formato: Materiale a stampa
• Livello bibliografico: Monografia
• Note generali: Description based upon print version of record.
• Nota di bibliografia: Includes bibliographical references and index.
• Nota di contenuto: Cover; Multidisciplinary Design Optimization in Computational Mechanics; Title Page; Copyright Page; Table of Contents; Foreword; Notes for Instructors; Acknowledgements; Chapter 1. Multilevel Multidisciplinary Optimization in Airplane Design; 1.1. Introduction; 1.2. Overview of the traditional airplane design process and expected MDO contributions; 1.3. First step toward MDO: local dimensioning by mathematical optimization; 1.4. Second step toward MDO: multilevel multidisciplinary dimensioning; 1.5. Elements of an MDO process; 1.6. Choice of optimizers; 1.6.1. Deterministic algorithms; 1.6.2. Stochastic algorithms1.7. Coupling between levels; 1.7.1. Reduction of mathematical models; 1.7.2. Simplified physical models; 1.8. Post-processing; 1.8.1. Lagrange multipliers; 1.8.2. Pareto fronts; 1.8.3. Self-organizing maps; 1.9. Conclusion; Chapter 2. Response Surface Methodology and Reduced Order Models; 2.1. Introduction; 2.2. Introducing some more notations; 2.3. Linear regression; 2.3.1. Introduction to linear regression; 2.3.2. Leverage; 2.3.3. Generalized linear regression; 2.3.4. An implicit reduced order model: moving least-squares (MLS) method; 2.3.5. Bias-variance trade-off2.4. Non-linear regression; 2.4.1. Neural networks as an example of non-linear models; 2.4.2. Another example of a non-linear model: parametrized RBFs; 2.4.3. Gradient algorithms; 2.4.4. Second-order methods; 2.5. Kriging interpolation; 2.5.1. Recall on Gaussian regression; 2.5.2. Basic principles of kriging algorithms; 2.5.3. Trend estimation; 2.5.4. Covariance estimation; 2.6. Non-parametric regression and kernel-based methods; 2.6.1. Introduction to non-parametric methods; 2.6.2. Parzen window regression; 2.6.3. Radial basis functions (RBFs); 2.6.4. EM estimation of a mixture2.6.5. How RBFs are used in this book; 2.7. Support vector regression; 2.7.1. Variational formulation of SVR; 2.7.2. Dual formulation of SVR; 2.7.3. Computation of SVR models; 2.7.4. Self-reproducing Hilbert space; 2.7.5. Regularizing properties of the kernel; 2.7.6. Margin selection and ν-regression; 2.7.7. Large databases and recursive learning; 2.8. Model selection; 2.8.1. Estimating generalization error; 2.8.2. Cross-validation methods; 2.8.3. Leverage methods; 2.9. Introduction to design of computer experiments (DoCE); 2.9.1. Classical techniques; 2.9.2. Input space sampling2.9.3. Adaptive learning and sequential design; 2.10. Bibliography; Chapter 3. PDE Metamodeling using Principal Component Analysis; 3.1. Principal component analysis (PCA); 3.2. Truncation rank and projector error; 3.3. Application: POD reduction of velocity fields in an engine combustion chamber; 3.4. Reduced-basis methods, numerical analysis; 3.4.1. POD-Galerkin projection method; 3.4.2. Dual approach POD-Petrov-Galerkin; 3.5. Intrusive/non-intrusive aspects; 3.6. Double reduction in both space and parameter dimensions; 3.7. The weighted residual method; 3.8. Non-linear problems
• Sommario/riassunto: This book provides a comprehensive introduction to the mathematical and algorithmic methods for the Multidisciplinary Design Optimization (MDO) of complex mechanical systems such as aircraft or car engines. We have focused on the presentation of strategies efficiently and economically managing the different levels of complexity in coupled disciplines (e.g. structure, fluid, thermal, acoustics, etc.), ranging from Reduced Order Models (ROM) to full-scale Finite Element (FE) or Finite Volume (FV) simulations. Particular focus is given to the uncertainty quantification and its impact on the robus

Graph Neural Network for Feature Extraction and Classification of Hyperspectral Remote Sensing Images / / by Yao Ding, Zhili Zhang, Haojie Hu, Fang He, Shuli Cheng, Yijun Zhang

• Record Nr.: UNINA9910910493103321
• Autore: Ding Yao
• Pubbl/distr/stampa: Singapore : , : Springer Nature Singapore : , : Imprint : Springer, , 2024
• ISBN: 9789819780099; 9819780098
• Edizione: [1st ed. 2024.]
• Descrizione fisica: 1 online resource (189 pages)
• Collana: Intelligent Perception and Information Processing, , 3059-3816
• Altri autori (Persone): ZhangZhili; HuHaojie; HeFang; ChengShuli; ZhangYijun
• Disciplina: 621.382
• Soggetti: Image processing; Neural networks (Computer science); Machine learning; Image Processing; Mathematical Models of Cognitive Processes and Neural Networks; Machine Learning
• Lingua di pubblicazione: Inglese
• Formato: Materiale a stampa
• Livello bibliografico: Monografia
• Nota di contenuto: Introduction -- Graph sample and aggregate-attention network for hyperspectral image classification -- Multi-feature fusion: Graph neural network and CNN combining for hyperspectral image classification -- Pixel and hyperpixel level feature combining for hyperspectral image classification -- Global dynamic graph optimization for hyperspectral image classification -- Exploring relationship between transformer and graph convolution for hyperspectral image classification.
• Sommario/riassunto: This book deals with hyperspectral image classification using graph neural network methods, focusing on classification model designing, graph information dissemination, and graph construction. In the book, various graph neural network based classifiers have been proposed for ; hyperspectral image classification to improve the classification accuracy. This book has promoted the application of graph neural network in hyperspectral image classification, providing reference for remote sensing image processing. It will be a useful reference for researchers in remote sensing image processing and image neural network design.