Global Optimization for hyper-parameter setting
Many optimization algorithm, as well as many calibration models or machine learning techniques, rely on optimization methods which, in turn, depend on the values of some parameters. Choosing these parameters can be considered as an optimization problem in which the objective function is a computationally expensive run of a specific numerical algorithm. The thesis will explore advanced global optimization algorithm for expensive objective functions. Simulation will most likely be performed using an high level language like python.