Bayes linear uncertainty analysis for complex physical systems modelled by computer simulators

Michael Goldstein (University of Durham)

Friday 13th March, 2015 15:00-16:00 Maths 204

Abstract

Most large and complex physical systems are studied by 
mathematical models, implemented as high dimensional computer 
simulators.  While all such cases differ in physical description, each 
analysis of a physical system based on a computer simulator involves 
the same underlying sources of uncertainty. There is a growing field 
of study which aims to quantify and synthesise all of the 
uncertainties involved in relating models to physical systems, within 
the framework of Bayesian statistics, and to use the resultant 
uncertainty specification to address problems of forecasting and 
decision making based on the application of these methods. This talk 
will give an overview of aspects of this emerging methodology, with 
particular emphasis on the Bayes linear approach to emulation, 
structural discrepancy modelling, iterative history matching and 
forecasting. The methodology will be illustrated with examples of 
current areas of practical application, and, in particular, to the 
analysis of flood models.

Add to your calendar

Download event information as iCalendar file (only this event)