CDS/CIMMS Lunchtime Seminar: Markovian approaches to approximate modeling for decision and control

Dr. Sean Meyn, Department of Electrical and Computer Engineering, University of Illinois at Urbana-Champaign

This talk begins with an overview of one class of techniques to construct low dimensional Markov models for complex systems. The origins of this approach are contained in Shannon's famous 1948 paper A mathematical theory of communication, which is regarded as the birth of modern information theory.  The technique was discovered independently in the physics community by Mori and Zwanzig in the 1960s.

New approaches to modeling will be described that combine elements of Shannon's construction with recent techniques from the stability theory for Markov models. Applications to control and to inference will be surveyed.