Hierarchical and Distributed Model Predictive Control, Algorithms and Applications

Moritz Diehl
Electrical Engineering Department (ESAT-SCD),
K.U.Leuven, Kasteelpark Arenberg 10,
3001 Leuven, Belgium
Phone: +32-16-321884 (Secr: -321122/-321709, Fax: -328539)

Riccardo Scattolini
Dipartimento di Elettronica e Informazione
Politecnico di Milano
Via Ponzio 34/1
20133 Milano
Tel. +39-02-23993539
Fax. +39-02-23993412
e-mail: riccardo.scattolini@elet.polimi.it

one full day (6 hours)

Sunday, August 28, 2011

Goal of the workshop
Aim is to present recent advances on hierarchical and distributed model predictive control, with the presentation of significant case studies

The workshop is aimed to present recent advances in the field of hierarchical and distributed control and estimation for large-scale complex networked systems. The main technique underlying all the proposed solutions is Model Predictive Control, in view of its flexibility in the definition of the control problem and of the possibility to include in the problem formulation state and control constraints.
Two mainstreams of recent research in the field will be covered. The first one refers to distributed optimization techniques for the solution of a centralized MPC problem. In this case, the goal is to decompose the optimization problem into a number of smaller and more easily tractable ones. In this framework, primal and dual approaches will be considered. The second approach relies on the solution of a number of local control problems with information exchange among them. In this case, the control algorithm itself, rather than its numerical solution, is distributed. Convergence properties of the methods can be achieved by resorting to robust MPC algorithms, where the uncertainties are related to the mutual influences among the subsystems. In the same way, it will be shown how to construct hierarchical control methods, where the hierarchical structure stems either from a structural decomposition of the system under control, or from its multi-level and multi time scale description.
A number of examples will be discussed to witness the potentialities of the methods. In particular, reference will be made to spatially distributed systems, such as irrigation channels and water networks. A complex application will deal with the control of a hydroelectric power valley, with five reservoirs, three river reaches and a number of additional plants (ducts, turbines, generators, dams). The design of a hierarchical control scheme for Combined Cycle Power Plants will also be discussed, with particular emphasis to the problems related to the start-up phase, where particular attention must be posed to the thermal and mechanical stresses of the components, which strongly affect the life time of the plant.


M. Diehl, R. Scattolini

J. Rawlings
University of Winsconsin, USA
An overview of distributed MPC

M. Diehl, A. Kozma, C. Savorgnan
KU Leuven, Belgium
Hierarchical and Distributed Optimization Algorithms

M. Farina, B. Picasso, R. Scattolini
Politecnico di Milano, Italy
Design of hierarchical and distributed MPC control systems with robustness tools

B. De Schutter
TUDelft, The Netherlands
Hierachical MPC with applications in transportation  and infrastructure networks

D. Muñoz de la Peña, J.M. Maestre and D. Limón
University of Seville, Spain
Distributed MPC based on game theory

W. Marquardt, H. Scheu
RWTH, Aachen, Germany
Distributed model predictive control by primal decomposition

D. Faille, F. Davelaar
Hierarchical and distributed control of a hydro power valley

A.Tica, H Guéguen, D. Dumur, D. Faille, F Davelaar
Supelec and EDF, France
Application to start-up of Combined-Cycle Power Plant

L. Sanchez, M. A. Ridao (USE)
INOCSA and University of Seville, Spain
Distributed Control of Irrigation Canals

B. De Schutter
TUDelft, The Netherlands

Riccardo Scattolini, Moritz Diehl
Sunday August 28, 2011