The goal of this proposal is to create mathematical and statistical theory for Extreme Episodes in random processes, and to use the theory to solve significant practical problems. The project will revolve around two mathematical themes: High-dimensional and detailed statistical models and methods for extreme episodes; and the *shape* of extreme episodes in non-differentiable Gaussian processes. Science and technology are producing enormously large data sets, at an ever increasing rate. These data sets are ready to be used for scientific discovery and technological advance, in ways beyond present imagination. The theory of extreme episodes will be one part of the answer to this challenge. What is needed is a much more sophisticated understanding of extreme behavior than what is given by existing theory. The results of this project will aid efforts to mitigate the impact of the extreme floods, windstorms, and heat waves which might be caused by a changing climate. They will lead to improved methods to avoid both fatal and less serious car accidents, and they will be used to diminish financial risks.
Professor vid Chalmers University of Technology, Mathematical Sciences, Applied Mathematics and Statistics
Funding Chalmers participation during 2012–2014