CS757/CS857: Mathematical Optimization for Applications

Co-taught with Mark Lyon.

When and Where

Mon & Wed, 9:40 am - 11:00 am Wed: Recitation

Class Content

The goal of the class is to teach the foundations that underlie mathematical optimization techniques. The methods play an important role in machine learning, operations research, applied mathematics, industrial engineering.

  1. Unconstrained optimization methods
  2. Gradient descent and line search
  3. Trust region methods
  4. Newton and Quasi-Newton methods (BFGS)
  5. Constrained optimization problems
  6. Lagrange multiplies and KKT conditions
  7. Linear and quadratic programming
  8. Convex analysis

Programming Language

MATLAB, Python, R, or Julia

Pre-requisites

  • Calculus
  • Some linear algebra background

Textbook

Nocedal and Wright, Numerical Optimization (2nd edition), 2006