A basic optimization course (such as LINMA1702) and basic knowledge in real analysis and linear algebra (such as provided by FSAB1101 and FSAB1102)
Linear optimization, convex optimization (including structured conic optimization) ; duality and applications ; interior-point methods ; first-order methods ; trust-region methods ; use of a modeling language.
The contribution of this Teaching Unit to the development and
command of the skills and learning outcomes of the programme(s) can be
accessed at the end of this sheet, in the section entitled
“Programmes/courses offering this Teaching Unit”.
At the end of this learning unit, the student is able to :
AA1.1, AA1.2, AA1.3
AA2.1, AA2.2, AA2.4, AA2.5
More specifically, at the end of the course the student will be able to :
recognize the possiblity of formulating or converting a problem into a linear, convex or conic optimization program
exploit the concept of duality in order to understand a problem, produce optimality or impossibility certificates, carry out sensitivity analysis or formulate robust problems
describe, analyze and implement advanced algorithms to solve linear, convex or non-linear optimization problems
use a modeling language to formulate and solve optimization problems, while understanding and exploiting the formal separation between model, data and resolution algorithm
Transversal learning outcomes :
use a numerical/computational software tool such as MATLAB, or a modeling language such as AMPL
formulate, analyze and solve optimization models, in a small group
write a report about the formulation, analysis and resolution of optimization models, in a small group
Models: Advanced modeling techniques for linear and convex optimization ; structured conic optimization ; convex duality with applications (alternatives, sensitivity analysis and robust optimization) ; Lagrangian duality
Methods: path-following interior-point methods for convex optimization (self-concordant barriers) ; first-order methods for convex and non-convex optimization (including random methods) ; algorithmic complexity and convergence rates ; trust-region methods ; introduction to the AMPL modeling language.
Applications in various domains, such as data analysis, machine learning, finance, shape or structural optimization (mechanics), telecommunications, etc.
The course is comprised of lectures, exercise sessions and computer labs, as well as a series of homeworks to be carried out in small groups.
Students will be evaluated with an individual written exam, based on the above-mentioned objectives. Students also carry out a series of homeworks in small groups, which are taken into account for the final grade.
Convex Optimization, Stephen Boyd et Lieven Vandenberghe, Cambridge University Press, 2004.
Lectures on Modern Convex Optimization: Analysis, Algorithms, and Engineering Applications, Aharon Ben-Tal, Arkadi Nemirovski, SIAM 2001.
Interior point methods for linear optimization, Cornelis Roos, Tamas Terlaky, Jean-Philippe Vial, Springer, 2006.
Introductory Lectures on Convex Optimization: A Basic Course, Yurii Nesterov, Kluwer, 2004.
Trust-region methods, A. Andrew R. Conn, Nicholas I. M. Gould, Ph. Philippe L. Toint, SIAM, 2000.
Title of the programme
Master  in Data Science Engineering
Master  in Computer Science and Engineering
Master  in Biomedical Engineering
Master  in Statistic: General
Master  in Computer Science
Master  in Mathematical Engineering
Master  in data Science: Information technology