Mechanical Engineering Seminar: Dr. J Tinsley Oden

Dr. John Tinsley Oden will be presenting a special seminar that is open to the public.

All dates for this event occur in the past.

E001 Scott Lab
201 W. 19th Ave.
Columbus, OH 43210
United States

Title:

FOUNDATIONAL ISSUES IN COMPUTATIONAL SCIENCE: Model Validation, Error control, Multiscale Modeling, and Uncertainty Quantification

Abstract:

The beginnings of a revolution in computational science is under way that could change the way computer models are perceived and implemented in the future. It has to do with what is termed “predictive modeling” – the systematic recognition of model and data uncertainties and their propagation through a computational model to produce predictions of quantities of interest with quantifiable uncertainty. Predictive modeling includes the processes of model calibration, validation, code and solution verification, estimation and control of error due to the choice of the model, as well as the discretization of the model, and the development of reasonable metrics for measuring uncertainty, quantifying it, and systematically reducing it when feasible. This emerging view of modeling and simulation is increasingly important as computer predictions are relied upon more frequently to make critical decisions affecting humankind – for example, in models of climate change, energy systems, biological systems, drug design, natural hazards, medical procedures, and nano-manufacturing, etc. In this lecture, we describe a general framework for model validation, a posteriori estimation of modeling and discretization error, and uncertainty quantification based on Bayesian statistics. Included are ideas on model selection and upgrading based on versions of evidence theory. We also describe a theory of adaptive modeling and demonstrate its application to rigorous approaches to multiscale modeling of systems that bridge molecular-to-continuum scales. Among applications of these approaches discussed are adaptive modeling of polymer components in nano-manufacturing and model validation and selection in predicting tumor growth.

Bio:

Dr. John Tinsley Oden is the founding Director of ICES (www.ices.utexas.edu), which was created in 2003 as an expansion of the Texas Institute for Computational and Applied Mathematics, also directed by Oden for over a decade. The Institute supports broad interdisciplinary research and academic programs in computational engineering and sciences. He is a member of; the U.S. National Academy of Engineering, the American Academy of Arts and Sciences, and the National Academies of Engineering of Mexico and of Brazil. He serves on numerous organizational, scientific, and advisory committees for international conferences and symposiums. He is an Editor of Computer Methods in Applied Mechanics and Engineering and serves on the editorial board of 24 scientific journals. He has worked extensively on the mathematical theory and implementation of numerical methods applied to problems in solid and fluid mechanics and, particularly, nonlinear continuum mechanics. He is the author or editor of over 500 scientific works, including 50 books and monographs. The Institute for Scientific Information lists Dr. Oden as one of the most highly cited researchers in the world in refereed, peer-reviewed journals. His current research focuses on the development of computational methods for multi-scale modeling, with applications to semi-conductor manufacturing and on computer models for the adaptive control of laser therapies for cancer.

For more information, contact Somnath Ghosh (.5, 2 2599) or J.K. Lee (.71, 2 7371)