Lecture: “Model Predictive Control: The Impact of Computation on Control”
by Prof. Manfred Morari
Recorded on August 30, 2019 at Seoul National University Library, Yang Doo Suk Hall
Since the early days of control, the complexity of control algorithms has been constrained by the hardware available for the on-line implementation. The arrival of control computers in the 1960s changed this with Model Predictive Control as the first algorithm to take advantage of the ever-growing computational resources. I will trace these developments and describe their impact on a wide range of areas from building climate control to automotive systems and power electronics. I will speculate on the future role of the control field, the potential impact of artificial intelligence, current research activities and public perception.
Satellite Lecture: Model Predictive Control
Date & Time: August 31 (Sat), 2019, 9:30 AM ~ 1:00 PM
Place: Seoul National University Library, Kwanjeong Building, Yang Doo Suk Hall
The basic idea of MPC is to solve a trajectory optimization problem on-line in real time and to apply the computed control moves in a receding horizon fashion. In this manner one can address complex control tasks involving a large number of inputs and outputs including constraints and even deal with nonlinear and hybrid (switched) systems. MPC does, however, require a model of the systems to be controlled and sufficient computational resources to solve the optimization problem in real time. On the other hand, MPC avoids the need to solve the Bellman equation. Much of the course is based on the recent book by Borrelli, Bemporad and Morari. We start by explaining the relationship with the classic LQR to illustrate the fundamental principles. We derive the recursive feasibility and stability conditions. We cover computational aspects. We will introduce a series of advanced topics like robust MPC, adaptive/learning MPC and MPC for hybrid systems. We will illustrate the techniques with a series of examples from academic labs and industry. “Model Predictive Control for Linear and Hybrid Systems” by Borrelli, Bemporad and Morari, Cambridge University Press (2017) https://goo.gl/fTYXSw
Handout download: MPC_Tutorial.pdf
About the speaker
In 2016 Manfred Morari was appointed Distinguished Faculty Fellow in the Department of Electrical and Systems Engineering at the University of Pennsylvania. In the prior 22 years he was a professor at ETH Zurich, from 2009 to 2012 head of the Department of Information Technology and Electrical Engineering and from 1994 to 2008 head of the Automatic Control Laboratory. Before that he was the McCollum-Corcoran Professor of Chemical Engineering and Executive Officer for Control and Dynamical Systems at the California Institute of Technology. From 1977 to 1983 he was on the faculty of the University of Wisconsin. He obtained the diploma from ETH Zurich and the Ph.D. from the University of Minnesota, both in chemical engineering.
Morari was president of the European Control Association 2011-2013 and General Chair of the European Control Conference in 2013.
Morari supervised more than 80 PhD students who are now professors in universities all over the world or hold positions of major responsibility in industry.
Throughout his career, Morari and his research group were striving to develop theoretically founded tools for process systems engineering, control and automation to further understanding and to impact practice. He co-authored the books “Robust Process Control” with Zafiriou (1989) and “Predictive Control for Linear and Hybrid Systems” with Borrelli and Bemporad (2017)
In the 1980s Morari’s research focused on robust control. The principles of Internal Model Control (IMC) introduced by him have been incorporated in all the leading process control textbooks and have been adopted by a number of control vendors and major users. The insights from this work helped Morari gain a deeper understanding of the nature of control and helped him establish the foundation for a systematic method to include operability aspects in chemical process design.
In the 1990s Morari became interested in the control of systems involving logic and constraints. Together with Bemporad he introduced a new modeling paradigm for linear hybrid systems, Mixed Logic Dynamical (MLD) Systems. Taking advantage of the rapidly increasing computer power, they showed how Model Predictive Control based on MLD system models can solve a range of challenging control problems in many applications ranging from building climate control to automotive and power electronic systems. ABB’s cement mill and electrical drive control systems are based on this technology.
Morari’s research is internationally recognized. The analysis techniques and software developed in his group are used in universities and industry throughout the world. He has received numerous awards, including the Eckman Award, Ragazzini Award and Bellman Control Heritage Award from the American Automatic Control Council; the Colburn Award, Professional Progress Award and CAST Division Award from the American Institute of Chemical Engineers; the Control Systems Technical Field Award and the Bode Lecture Prize from IEEE; the Nyquist Lectureship and the Rufus Oldenburger Medal from the American Society of Mechanical Engineers; the IFAC High Impact Paper Award for his publication “Control of systems integrating logic, dynamics and constraints” co-authored by A. Bemporad. He is a Fellow of IEEE, AIChE and IFAC. In 1993 he was elected to the U.S. National Academy of Engineering and in 2015 to the UK Royal Academy of Engineering. Manfred Morari served on the technical advisory boards of several major corporations.