Course on LMI optimization with applications in control
(Didier Henrion, LAAS-CNRS, Toulouse, France and FEL-CVUT, Prague,
Czech Republic)
Czech Technical University,
Prague, Czech Republic - April 2013
Venue and dates
The course is given at the Charles Square campus of the
Czech Technical University, in the historical center of
Prague (Karlovo Namesti 13, 12135
Praha 2) during the first week of April 2013.
It consists of seven two-hour lectures, given on Tuesday 2 April,
Wednesday 3 April, Thrusday 4 April from 10am to noon
and from 2pm to 4pm, and Friday 5 April from 10am to noon.
The course is given in room K14,
ground floor, to your left handside when entering building E.
Please refer to these maps
for instructions to reach this building. The campus entrance
is restricted to badge holders, so please ask the doorperson
to let you in through the electronic gates.
Registration
The course is primarily aimed at students from the Czech Technical
University in Prague, yet external participants are welcome. There is
no registration fee. Please note that the Czech Technical University
will not provide assistance regarding traveling and accomodation in
Prague.
Description
This is a course for graduate students or researchers
with a background in linear control
systems, linear algebra and convex optimization.
The focus is on semidefinite programming (SDP), or
optimization over linear matrix inequalities (LMIs), an extension of
linear programming to the cone of positive semidefinite matrices. Since
the 1990s, LMI methods have found numerous applications mostly in
combinatorial optimization, systems control and signal processing.
Outline
The course starts with fundamental
mathematical features of linear matrix inequalities:
Part 0: general introduction, course outline and material
Part 1: technical background on linear algebra, numerical
methods, cones, duality and convexity
Part 2: what is an LMI ? history, connection with semidefinite
programming (SDP)
Part 3: LMI duality, alternatives, Farkas lemma, S-procedure
More recent material covers the applications of LMIs
in polynomial optimization:
Part 4: convex LMI modelling,
classification of convex semialgebraic sets that can
be represented with LMIs, lift and project techniques
Part 5: nonconvex LMI modelling, BMI,
Shor's relaxation, polynomial moments and sum-of-squares,
Lasserre's hierarchy of LMI relaxations for non-convex polynomial
optimization
We also briefly survey recent developments in semidefinite solvers and
software packages:
Part 6: LMI solvers, basics of interior-point algorithms,
latest achievements in software and solvers for LMIs and BMIs
The end of the course focuses on the use of measures
for static polynomial optimisation problems and occupation
measures for differential equations and related optimal
control problems:
Part 7: Measures, occupation measures
and control problems
This last part surveys most of the material covered in Parts 5,
but through the lense of measure theory and the related
generalized problem of moments. Moreover, it extends
these ideas to differential equations and optimal control framework,
and it describes several research directions along these lines.
Homeworks and exam
Homeworks are handed out during the course.
A written examination can be organized.
Last updated on 27 March 2013.