The advances in LP and MIP solver APIs during the last 30 years have made implementing column generation algorithms more accessible. However, the algorithm has some counterintuitive properties that can make it seem mysterious, such as reformulating a compact model into one with a huge number of variables that cannot all be stored in available computer memory.

This informative webinar is hosted by Dr. Ed Klotz, Senior Mathematical Optimization Specialist at Gurobi, and Dr. Greg Glockner, Vice President and Technical Fellow at Gurobi. This webinar will try to provide an intuitive motivation for column generation to make the subsequent explanation of the underlying mathematics more accessible. While the presentation is primarily intended for those who are either unfamiliar with the concepts or find them challenging when exposed to them, those more familiar with column generation methods may still gain some new perspectives as well.