General

3 Days Paid Course on Tractable Techniques for Robust and Explicit Model Predictive Control at IIT Bombay: Register by June 9

Apoorva

May 22, 2023
3 min read

0 Comments

Online Applications are invited for three days paid course on Tractable Techniques for Robust and Explicit Model Predictive Control at IIT Bombay. The last to register is June 9, 2023.

About IIT Bombay

IIT Bombay was established in 1958, the second of its kind, IIT Bombay was the first to be set up with foreign assistance. The funds from UNESCO came as Roubles from the then-Soviet Union. In 1961 Parliament decreed the IITs as ‘Institutes of National Importance’. Since then, IITB has grown from strength to strength to emerge as one of the top technical universities in the world.

Course Details

Model predictive control (MPC) is one of the most applicable control techniques available today; it has been widely adopted by the industry in general and has witnessed explosive growth and proliferation in the process industries in particular.
It is an optimization-based method designed for implicit feedback control of systems. It takes into account all constraints at the design stage and simultaneously adjusts all control actions applied to multi-input multi-output systems while accounting for all interactions within the systems.
This leads to minimal conservatism and aggressive control of the underlying processes close to the admissible boundaries of operation while adhering to strict safety standards.
MPC has witnessed intense activity over the past several decades, and both the theoretical and practical aspects of this technique are reasonably well-developed, especially in the linear regime.
However, nonlinear models and uncertain environments pose significant challenges to the numerical tractability front for MPC.
Specifically, numerically tractable algorithms for MPC in the nonlinear regime are rare.
Moreover, explicit MPC of nonlinear systems and systems with uncertainty is a challenging topic fraught with stiff challenges.
This course is designed to expose entirely novel ideas, with emphasis on approximation-free computationally tractable methods, for explicit MPC for both nonlinear systems and systems with uncertain models.

Course Content

The program will have lectures and tutorials which will give hands-on experience to the participants with regard to solving problems and performing calculations.
Basics of MPC (2 hr): elementary theory of nominal MPC, effects of uncertainty, robust MPC, the ingredients of convexity; numerical exercises + solutions.
Techniques for uncertain and constrained optimal control (4 hr): a new technique for solving minimax problems in the convex regime, relationship with uncertain and constrained optimal control, global optimization, examples; numerical exercises + solutions.
Quasi-interpolation (2 hr): introduction to quasi-interpolation; exercises + solutions.
Explicit MPC, novel techniques (7 hr): quasi-interpolation driven explicit MPC; exercises + solutions.

Important Dates

Course Sessions will be conducted from June 19, 2023, to June 21, 2023
Registration Last Date: June 9, 2023.

Who May Benefit?

Students and professionals in Engineering working in the broad areas of Control, Process Control, and Robotics, who employ model predictive control in their work.
Students and professionals are interested in new results on various aspects of robust, stochastic, and explicit MPC.
The course will specifically emphasize novel numerical techniques for the solution of MPC problems; MATLAB/python/Mathematica programs will be extensively employed as illustrative examples during the course.