Heuristic Mixed-Integer Programming

Catalog description

Discrete Optimization problems are generally formulated through Mixed-Integer Linear and Nonlinear Programming (MIP) and solved, to a large extent, by both commercial and noncommercial MIP solvers. Solving the formulated models to proven optimality is generally not necessary on the application side and often out of reach. However, computing high quality feasible solutions is a strong requirement in applications. The course covers the methodology behind the development of effective heuristic methods within MIP solvers and presents recent hybridization mechanisms based on data and Machine Learning.