Thesis title: A Lipschizian Global Optimization Algorithm and Machine Learning for Fluid Dynamics
In this thesis, we propose a new algorithm for Global Optimization and Machine Learning based methodologies for Fluid Dynamics. In particular, in the first chapter, we will discuss how estimates of the Lipschitz constant obtained from the Extreme Value Theory of Statistics can be used inside a partition scheme for Global Optimization. In the second, third, and fourth chapters we will discuss how methodologies from Machine Learning field can be used in various applications of Fluid Dynamics. Specifically, in the second chapter, we will discuss a new framework for design space dimensionality reduction for shape optimization. In the third chapter, we will continue to use Unsupervised Learning methods in order to extract and interpret highly nonlinear data obtained from Particle Image Velocimetry measurements of turbulent phenomena. In the last chapter, we will show the application of some recent Supervised Machine Learning methods for ship motion prediction in a high sea state level.