Non-Convex Quadratic Optimization

Webinar Summary

One major new feature in Gurobi 9.0 is a new bilinear solver, which allows users to solve problems with non-convex quadratic objectives and constraints (i.e., QPs, QCPs, MIQPs, and MIQCPs). Many non-linear optimization solvers search for locally optimal solutions to these problems.

In contrast, Gurobi can now solve these problems to global optimality. Non-convex quadratic optimization problems arise in various industrial applications. In particular, non-convex quadratic constraints are vital to solve classical pooling and blending problems.

In this webinar session, we will:

• Introduce MIQCPs and mixed-integer bilinear programming
• Discuss algorithmic ideas for handling bilinear constraints
• Show a live demo of how Gurobi 9.0 supports bilinear constraints by building and solving a small instance of the pooling problem

Presented Materials

You can download the PDF with the slides here and the pooling problem Jupyter Notebook here.