Statistical Solutions in Physical and Computational Experiments

• Activity Number: 99
• Type: Invited
• Date/Time: Monday, August 4, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Section on Physical and Engineering Sciences
• Author(s): Leslie M. Moore*+
• Companies: Los Alamos National Laboratory
• Keywords: experiment design ; computer experiments ; multi-level experiment ; linear trend ; orthogonal array ; Latin hypercube sampling

Abstract

Statistical research is often required to address challenges posed by both physical and computer experiments. Statistical methodology motivated by physical and computational experiments is described to address statistical issues associated with materials experiments and infrastructure simulations. Materials experiments have stimulated statistical developments associated with multi-level experiments, presence of a linear trend, and experimental plans to address constraints. Similarly, infrastructure simulations have stimulated development of Latin hypercube sampling techniques, OA-based LHS designs, and other statistical innovations such as calibration and uncertainty quantification. These methodological advances will be described and illustrated in the context of physical materials experiments and infrastructure simulation studies.