Class 9

We continue studying decision-making problems under uncertainty with recourse. We show how this type of problems can be, in some cases, addressed by adjustable robust optimization. We focus on polyhedral uncertainty sets (independent of the here-and-now decision variables) and on recourse problems where uncertainty affects the constraint matrix or the right-hand-side, so that the worst-case recourse cost can be computed by exploring the vertexes of the polyhedral uncertainty set.

Adjustable robust optimization is exemplified by solving a robust version of the economic dispatch problem that accounts for wind generation uncertainty and that allows for re-dispatch decisions in real time. This teaching and learning activity has been prepared by one of the groups, The Unconstrained.