Modern Algorithmic Optimization

April 2 – 13, 2018

Speaker: Yury Nesterov (Université catholique de Louvain, Belgium)

TIn this course we present the most important research directions in the modern Optimization Theory. The main topics of our interest are related to the provable complexity of optimization problems and the most efficient methods for finding their approximate solution. The main attention will be given to the methods for solving problems of large and super-large dimension, which arise in many engineering applications, telecommunications, and models for analyzing the Internet activity. We consider also the optimization schemes, which are necessary for justifying rationality of consumers in economic models. The most of the material is absent in the monographic literature. Therefore, we include in the course all necessary proofs. The course will be giving in two weeks, April 2-6 and 9-13 (2017) by ten lectures, two hours each. During the course, we are going to discuss the following topics.

Topic 1. Complexity of optimization problems
Topic 2. Universal first-order methods
Topic 3. Second-order methods. Systems of nonlinear equations
Topic 4. Looking into the Black Box: Smoothing technique
Topic 5. Looking into the Black Box: Interior-point methods
Topic 6. Optimization with relative accuracy
Topic 7. Solving the huge-scale optimization problems
Topic 8. Algorithmic models of human behavior

Venue: The mini-course will be held at the HSE faculty of Computer Science (3 Kochnovsky Proezd building 1):
Mondays – room 402
Tuesdays – room 509
Wednesdays – room 509
Thursdays – room 205
Fridays – room 402

Time:  10:00 – 12:00