Gaussian Processes for SOLPS Data Emulation

Computer codes modeling plasma-wall interactions of fusion plasmas are costly in computer power and time—the running time for a single parameter setting is easily on the order of weeks or months, not to mention the expenditure for parametric studies. We propose to exploit the already gathered results in order to predict the outcome in the high-dimensional parameter space. For this, we utilize the Gaussian process method within the Bayesian framework. Uncertainties of the predictions are provided that point the way to parameter settings of further (expensive) simulations.