Skip to content Skip to main navigation Report an accessibility issue

A Duality Framework for Online Optimal Control in Transportation Systems

MalikopoulosDr. Andreas A. Malikopoulos
R&D Staff, Alvin M. Weinberg Fellow
Energy and Transportation Science Division, Oak Ridge National Laboratory
November 8, 2013, 2:30 – 3:30 PM
500 John D. Tickle Engineering Building

Dr. Andreas A. Malikopoulos received a Diploma in Mechanical Engineering from the National Technical University of Athens, Greece, in 2000. He received M.S. and Ph.D. degrees from the Department of Mechanical Engineering at the University of Michigan, Ann Arbor, in 2004 and 2008, respectively. His research interests span several fields, including analysis, optimization, and control of stochastic systems; stochastic optimal control; nonlinear optimization and convex analysis; large-scale optimization; and learning in complex systems. The emphasis is on applications related to energy, intelligent transportation, and operations research. Before joining Oak Ridge National Laboratory, he was a Senior Researcher with General Motors Global Research & Development, conducting research in the areas of stochastic optimization and control of advanced propulsion systems. He has been selected by the National Academy of Engineering to participate at the annual 2010 German-American Frontiers of Engineering Symposium, and the 2012 NAKFI conference, The Informed Brain in a Digital World.

Talk Abstract: The increasing necessity for environmentally conscious vehicle designs, in conjunction with increasing concerns regarding US dependency on foreign oil, has led to significant enhancement of the transportation portfolio with new vehicle technologies. Hybrid electric vehicles (HEVs) have attracted considerable attention due to their potential to reduce petroleum consumption and greenhouse gas emissions. A typical HEV consists of various subsystems, e.g., the internal combustion engine, the battery, and the electric machines (motor and generator). Implementing online a power management control policy to distribute the power demanded by the driver optimally to the available subsystems constitutes a challenging control problem and has been the object of intense study for the last two decades. This talk will address the development of a theoretical framework that can be used online to derive the optimal control policy for any different driver. The stochastic control problem is treated as a multiobjective optimization problem of the one-stage expected costs of the subsystems, and it is shown that the control policy yielding the Pareto efficient solution is an optimal control policy. The talk will conclude with highlighting current research efforts towards making intelligent vehicles with the aim of (1) becoming eco-friendly, (2) realizing the optimum performance and efficiency based on consumers’ needs and preferences, and (3) learning how traffic information can positively impact the environment.