"Optimal Control Under Uncertainties and Hard Constraints in Dynamic Nonlinear Systems"
Satyakiran Munaga ESAT-INSYS IMEC
Abstract:
In this talk we focus on the problem of dynamically calculating the
discrete trajectory of control variables towards the goal of optimizing
given operational cost metrics (e.g. energy or fuel consumption,
revenue generated, etc). In the systems of interest the uncertainties
influence not just the cost metrics but also the constraints (e.g.
deadline, temperature, lifetime, etc). Conventional approaches
guarantee the constraints based on the design-time worst-case bounds on
the uncertainties. In highly dynamic systems such bounds are very loose
and results in highly sub-optimal systems, if feasible at all. Use of
design-time scenarios improves the optimality but quickly explodes in
the presence of multiple sources of uncertainties. While the
nonlinearities in the system encourage predictive control, doing it
optimally is tricky due to the presence of interfering constraints. To
bridge the optimality gap with respect to the desired but infeasible
retrospective solution, we propose proactive control with the combined
strategy of dynamic bounding and dynamic likely-case optimization. We
also introduce the notion of and motivate the need for contextual turbo
modes. We illustrate the proposed concepts using a dynamic task
scheduling example of a realistic design where the proactive control
reduces the energy consumption up to 50% compared to state of the art
solutions.
Two OPTEC professors have been awarded three "Gouden Krijtjes", the yearly teaching awards given by the organization of engineering students (vtk). Prof. Lombaert was awarded the prize for the best course in civil engineering, and Prof. Diehl the prizes for the best professor and the best course in mathematical engineering (where he teaches numerical optimization). They received these awards at the yearly "proffentap" where experienced students taught them how to draft beer professionally.
