Fall 2023 WMM9AM16 Efficient methods in optimization
Instructor: Anatoli Juditsky Bat. IMAG, #156, email: anatoli.juditsky@univ-grenoble-alpes.fr
Lectures take place on Tue and Wed 5pm-6:30pm, Ensimag H105 except H206 on 09/26, D208 on 10/03, and H102 on 12/12
Assignment deadlines:
HW1 Tuesday Nov. 7, 2023 HW2 Monday Jan. 29, 2024
Part I Optimization theory
Preliminaries. Complexity of Nonlinear Optimization
Convex sets. Theory of Linear Programming
Convex functions. Convex Programming and Lagrange Duality
From Linear to Conic Programming
Conic Quadratic and Semidefinite Programs
Part II Optimization methods
Gradient Descent and Newton Method. Around Newton Method
Complexity of Convex Programming. Ellipsoid Method
Methods for high-dimensional convex problems Mirror Descent algorithm
Polynomial-time Interior-Point Methods
Algorithms of Stochastic Optimization
Fundamentals: affine spaces, differentiable functions, symmetric matrices