Course description
It is organized for 4th and 5th year students of the FUPM streams.
The purpose of the course is to teach students modern methods of convex optimization and their application in solving problems of convex and non—convex optimization. Special emphasis will be placed on conic optimization, starting with linear programming, and moving on to more complex problems of conic-quadratic and semi-definite programming, as well as problems over non-symmetric cones. In close conjunction with the methods themselves, there are issues of modeling, i.e. the presentation of specific tasks in the standard form of a conical program. If this is not possible, different techniques are used to construct convex approximations (relaxations), which can be reduced to the form of a conical program. We will also consider some standard methods of non-convex optimization based on solving a sequence of convex problems, in particular, methods such as constraints and branches.
Instructors
Roland Hildebrand