Varying Coefficient Models and Design Choice for Bayes Linear Emulation of Complex Computer Models with Limited Model Evaluations

Computer models are widely used to help make decisions about real-world systems. As computer of large and complex systems can have long run-times high-dimensional input spaces, it is often necessary use emulation assess uncertainties in model output. This paper presents methodology for motivated by a example energy policy. The studied an economic investment electricity generation Great Britain. was select parameters government policy designed incentivize renewable technologies meet targets. Limited computing time meant that few runs the were available fit emulator. statistical developed therefore focused on accurately capturing uncertainty output arising from small number runs. A varying coefficient emulator proposed when extrapolating away To maximize available, this paired with criterion-based procedure design selection.