This spring school is the 9th edition organized by the GdR MACS, a research group in modeling, analysis and control of dynamical systems. The mission of this recurrent school is to provide high-level courses in the fields of automatic control and adjacent disciplines covered by the GdR MACS.

Convex optimization methods, including semidefinite programming (SDP) and linear matrix inequalities (LMIs), have shaken the field of automatic control in the last three decades.

These methods have become a standard tool for the analysis of dynamical systems described by differential and linear recurrent equations, or involving linear equations combined with so-called « complications » (non-linear/infinite-dimensional operators, uncertainties, hybrid systems). Recent research progress allow us to perform the analysis of such systems with either pessimistic approaches (upper bounds on expected performance) or optimistic approaches (lower bounds reached by worse-case scenario).

Furthermore with the help of relaxation techniques, including the moment-sums of squares hierarchies for polynomial optimization, these pessimistic and optimistic bounds can be sometimes refined and come together with convergence guarantees to the initial problem's solution. Besides, these analysis tools also allow us to synthesize feedback control laws, filters, observers, even to design closed loop control laws while relying on specific heuristics. A strong knowledge of these methods, as well as the related algorithmic frameworks in convex optimization, shall provide in the forthcoming years the possibility to establish top-notch results for complex systems and large-scale systems, in particular network systems and dynamical systems coupled with neural nets.

The goal of this school is to provide a detailed overview of existing methods but also to identify their limitations, in order to allow researchers, PhD candidates and Postdoctoral fellows from the field, to establish their best contributions for the advance of knowledge.

The lectures will be performed in English.

Registration and logistics

The MACS 2023 school is recognized in the context of permanent formation for CNRS researchers and has been jointly funded  by INS2I and INSMI. The registration is free for researchers employed by CNRS (both permanent and non-permanent researchers) who can be fully funded to join the school via the permanent formation from their delegation. For this, please pre-register and ask a convocation to Gülgün Alpan. Next, please forward your convocation and a signed mission order to your regional delegation to ask for your trip funding.

The registration fees include all meals and lodging at the Paul Langevin center in Aussois from 8 to 12 May:

PhD/Postdocs 400€

Private institutions 650€

CNRS agents Gratuit

Other public institutions 570€

Transport

A free shuttle service has been settled on Monday 8 May and Friday 12 May, between the SNCF train station Lyon Part-Dieu and centre Paul Langevin of Aussois, with an intermediate stop at Lyon St Exupery airport.

Departure from Lyon Part-Dieu at 9h30 - Lyon St Exupery airport

Lyon St Exupery airport 10h20 - arrival at Aussois 12h30

12 May 2023

Departure Aussois at 14h - Lyon St Exupery airport

Lyon St Exupery airport at 16h20 - arrival at Lyon Part Dieu 17h00

The closest train station from the centre Paul Langevin of Aussois is Modane.

There are trains between the stations of Paris Gare de Lyon - Modane, Chambéry - Modane et Lyon - Modane.

Invited speakers

Delphine Bresch Pietri, MINES ParisTech

Tobias Holicki, Université de Stuttgart

Laurent Lessard, Northeastern University

Swann Marx, CNRS LS2N

Corbinian Schlosser, CNRS LAAS

Matteo Tacchi, CNRS Gipsa-Lab

Giorgio Valmorbida, L2S

Hasnaa Zidani, INSA Rouen

Organization committee

Gülgün Alpan, Grenoble INP - UGA

Victor Magron, CNRS LAAS

Dimitri Peaucelle, CNRS LAAS