Preface |
About the Authors |
Foreword |
Prologue |
Fundamentals / Part I: |
Sampling Plans / 1: |
The 'Curse of Dimensionality' and How to Avoid It / 1.1: |
Physical versus Computational Experiments / 1.2: |
Designing Preliminary Experiments (Screening) / 1.3: |
Estimating the Distribution of Elementary Effects / 1.3.1: |
Designing a Sampling Plan / 1.4: |
Stratification / 1.4.1: |
Latin Squares and Random Latin Hypercubes / 1.4.2: |
Space-filling Latin Hypercubes / 1.4.3: |
Space-filling Subsets / 1.4.4: |
A Note on Harmonic Responses / 1.5: |
Some Pointers for Further Reading / 1.6: |
References |
Constructing a Surrogate / 2: |
The Modelling Process / 2.1: |
Stage One: Preparing the Data and Choosing a Modelling Approach / 2.1.1: |
Stage Two: Parameter Estimation and Training / 2.1.2: |
Stage Three: Model Testing / 2.1.3: |
Polynomial Models / 2.2: |
Example One: Aerofoil Drag / 2.2.1: |
Example Two: a Multimodal Testcase / 2.2.2: |
What About the k-variable Case? / 2.2.3: |
Radial Basis Function Models / 2.3: |
Fitting Noise-Free Data / 2.3.1: |
Radial Basis Function Models of Noisy Data / 2.3.2: |
Kriging / 2.4: |
Building the Kriging Model / 2.4.1: |
Kriging Prediction / 2.4.2: |
Support Vector Regression / 2.5: |
The Support Vector Predictor / 2.5.1: |
The Kernel Trick / 2.5.2: |
Finding the Support Vectors / 2.5.3: |
Finding [mu] / 2.5.4: |
Choosing C and [epsilon] / 2.5.5: |
Computing [epsilon]: v-SVR / 2.5.6: |
The Big(ger) Picture / 2.6: |
Exploring and Exploiting a Surrogate / 3: |
Searching the Surrogate / 3.1: |
Infill Criteria / 3.2: |
Prediction Based Exploitation / 3.2.1: |
Error Based Exploration / 3.2.2: |
Balanced Exploitation and Exploration / 3.2.3: |
Conditional Likelihood Approaches / 3.2.4: |
Other Methods / 3.2.5: |
Managing a Surrogate Based Optimization Process / 3.3: |
Which Surrogate for What Use? / 3.3.1: |
How Many Sample Plan and Infill Points? / 3.3.2: |
Convergence Criteria / 3.3.3: |
Search of the Vibration Isolator Geometry Feasibility Using Kriging Goal Seeking / 3.4: |
Advanced Concepts / Part II: |
Visualization / 4: |
Matrices of Contour Plots / 4.1: |
Nested Dimensions / 4.2: |
Reference |
Constraints / 5: |
Satisfaction of Constraints by Construction / 5.1: |
Penalty Functions / 5.2: |
Example Constrained Problem / 5.3: |
Using a Kriging Model of the Constraint Function / 5.3.1: |
Using a Kriging Model of the Objective Function / 5.3.2: |
Expected Improvement Based Approaches / 5.4: |
Expected Improvement With Simple Penalty Function / 5.4.1: |
Constrained Expected Improvement / 5.4.2: |
Missing Data / 5.5: |
Imputing Data for Infeasible Designs / 5.5.1: |
Design of a Helical Compression Spring Using Constrained Expected Improvement / 5.6: |
Summary / 5.7: |
Infill Criteria with Noisy Data / 6: |
Regressing Kriging / 6.1: |
Searching the Regression Model / 6.2: |
Re-Interpolation / 6.2.1: |
Re-Interpolation With Conditional Likelihood Approaches / 6.2.2: |
A Note on Matrix Ill-Conditioning / 6.3: |
Exploiting Gradient Information / 6.4: |
Obtaining Gradients / 7.1: |
Finite Differencing / 7.1.1: |
Complex Step Approximation / 7.1.2: |
Adjoint Methods and Algorithmic Differentiation / 7.1.3: |
Gradient-enhanced Modelling / 7.2: |
Hessian-enhanced Modelling / 7.3: |
Multi-fidelity Analysis / 7.4: |
Co-Kriging / 8.1: |
One-variable Demonstration / 8.2: |
Choosing X[subscript c] and X[subscript e] / 8.3: |
Multiple Design Objectives / 8.4: |
Pareto Optimization / 9.1: |
Multi-objective Expected Improvement / 9.2: |
Design of the Nowacki Cantilever Beam Using Multi-objective, Constrained Expected Improvement / 9.3: |
Design of a Helical Compression Spring Using Multi-objective, Constrained Expected Improvement / 9.4: |
Example Problems / 9.5: |
One-Variable Test Function / A.1: |
Branin Test Function / A.2: |
Aerofoil Design / A.3: |
The Nowacki Beam / A.4: |
Multi-objective, Constrained Optimal Design of a Helical Compression Spring / A.5: |
Novel Passive Vibration Isolator Feasibility / A.6: |
Index |