Many long-term planning problems for power systems and energy policy have significant uncertainty, but also opportunities to learn and adapt. Often, better designs and policies build in additional flexibility at a small incremental cost to allow adaptation to what is learned later. Traditional stochastic programming and stochastic dynamic programming methods suffer from the curse of dimensionality, and are not feasible on large-scale problems. An alternative class of methods, sometimes called Approximate Dynamic Programming, offers a computationally feasible means of solving large-scale multi-stage stochastic problems. I will describe the basic algorithm structure and typical variants, and demonstrate the method on a climate policy model. I will also show initial results of ADP in power systems applications.
Prof. Webster is an Assistant Professor of Engineering Systems, with a focus on energy and environmental systems. Prof. Webster specializes in risk analysis, uncertainty analysis, and decision-making under uncertainty. He has published numerous peer-reviewed articles in energy and environmental science, economics, and policy, and has served on several national and international panels, including the U.S. Climate Change Science Program. Current research projects include stochastic dynamic modeling of the electric power system focusing on the integration of intermittent renewable generation (NSF), modeling technological change as a stochastic process and implications for near-term R&D portfolios (NSF, DOE), and flexible air quality strategies under uncertainty using integrated economic/energy/chemistry regional models (NSF, EPA). Prof. Webster is active in several research centers at MIT, including the Center for Energy and Environmental Policy Research (CEEPR), the Joint Program on the Science and Policy of Global Change, and the MIT Energy Initiative. Prior to returning to MIT, Prof. Webster was an assistant professor of public policy in the Department of Public Policy at the University of North Carolina at Chapel Hill. He received a Ph.D. (2000) in Engineering Systems and a M.S. (1996) in Technology and Policy from MIT, and a B.S.E. (1988) in Computer Science and Engineering from the University of Pennsylvania.
Autor: Mort Webster (MIT)
Lugar: Aula de seminarios del IIT
Persona de contacto: Javier García
Añadir a agenda electrónica: