Statistical Modelling for Dynamic Computer Simulators

Abstract

The recent accelerated growth in the computing power has generated popularization of experimentation with computer models (or simulators) in various physical and engineering applications. Realistic computer models of complex phenomena can be very expensive and statistical surrogates are used for detailed investigation. In this talk, I will present Gaussian Process model based surrogates for emulating the simulator outputs. I will particularly focus on a computationally efficient statistical emulator for a large-scale dynamic computer simulator (i.e., simulator which gives time series outputs). I will illustrate the proposed methodology using several test functions and a real-life application.