14:30
Imperial College
December 4
Lecture Theatre 1, ACE Extension
This year Professor Daoutidis will speak on: New Paradigms for the Automated Solution of Complex Control and Optimization Problems.
This event is open to all, but please note that the Lecture Theatre has limited capacity .
Abstract:
This talk will focus on the classic open challenge of solving complex, large-scale control and optimization problems efficiently and in an automated fashion. Such problems arise in process design and control due to process or plant integration and in the context of enterprise-wide optimization. Despite their complexity, they typically have structure and sparsity, which lends itself naturally to decomposition-based solution approaches. Yet finding proper decompositions has been addressed largely based on intuition. We have developed methods based on modern network science that generate automatically high-quality decompositions that can be integrated directly in structured solution approaches. These methods rely on detecting latent block structures (communities and hierarchies) in suitable graph representations of control and optimization problems. I will describe the theoretical and algorithmic advances, and applications to plantwide control and integration of scheduling and dynamic optimization problems, along with an industrial implementation. I will also discuss newly developed methods, based on machine learning, for aiding the selection and implementation of solution algorithms and accelerating process control and optimization. Collectively, these results highlight how advances in data and computer science can enable novel methods to address complex process systems engineering problems, very much in the spirit of Professor Sargent’s overarching research philosophy and legacy.
Bio:
Prodromos Daoutidis is a College of Science and Engineering Distinguished Professor, Distinguished University Teacher, and holder of the Amundson Chair in the Department of Chemical Engineering and Materials Science at the University of Minnesota. He is also the Founding Director of the M.S. degree program in Data Science for Chemical Engineering and Materials Science and the Director of the “Data Driven Discovery and Engineering from Atoms to Processes” National Science Foundation Research Traineeship Program. He received a Diploma degree in Chemical Engineering from the Aristotle University of Thessaloniki, M.S.E. degrees in Chemical Engineering and Electrical Engineering: Systems from the University of Michigan, and a Ph.D. degree in Chemical Engineering from the University of Michigan. He is the recipient of the IChemE Hutchison Medal, the AIChE Sustainable Engineering Forum Research Award, the AIChE Computing in Chemical Engineering Award, the C.A. Floudas Award in Mathematical Optimization, the PSE Model Based Innovation Prize, and Best Paper Awards from the Journal of Process Control and Computers and Chemical Engineering. He has served as CAST Programming Chair and CAST Chair and is currently the President of FIPSE. Daoutidis is the Associate Editor for Process Systems Engineering in the AIChE Journal and an Associate Editor in the Journal of Process Control. He has co-authored 5 books and over 350 refereed papers and has supervised to completion 44 Ph.D. students and post-docs, 15 of whom currently hold academic positions. His current research is on control and optimization of complex and networked systems, and the engineering and economics of green ammonia for sustainable energy and agriculture applications.