Simulations play a key role in helping decision makers respond to complex societal problems, ranging from strategies for enhancing infrastructure resilience to natural hazards, to understanding financial markets, or managing air traffic. However, to be effective, individuals must select appropriate models and understand their limitations. This project conducts fundamental research to create a computationally efficient paradigm for model validation designed to guide decisions on model accuracy and validity by exploring families of models and choosing those most useful. The project aims to develop an integrated framework of model falsification and Bayesian model selection: initial falsification eliminates unsuitable models that do not fit measured data, Bayesian model class selection quantifies the confidence in the models, and final falsification selects the model classes with the most accurate simulations. New approaches to model falsification provide a cohesive framework that accommodates many types of dynamic response data.
Research efforts at Clarkson are centered on overcoming the computational challenges encountered when applying the proposed framework to realistic problems. This will be accomplished by exploring the use of state-of-the-art uncertainty quantification algorithms, including stochastic collocation, polynomial chaos, sparse grid quadrature techniques, and sparse collocation techniques to enhance the computational efficiency of model falsification and Bayesian model selection. Additionally, UQ algorithms for systems with local features, pioneered by my group and collaborators, will be investigated in model validation.
This project is in collaboration with E.A. Johnson [USC] and P.T. Brewick [US NRL] and supported by NSF CMMI-1662992 [Clarkson] & NSF CMMI-1663667 [USC].