Technical Note—Multistage Robust Mixed-Integer Programming

In the paper “Multistage Robust Mixed-Integer Programming,” Krzysztof Postek, Ward Romeijnders, and Wolfram Wiesemann study multistage robust mixed-integer programming problems, a popular and powerful yet computationally demanding framework for decision making under uncertainty. The authors leverage dynamic programming principles to decompose multistage problems into large numbers of two-stage subproblems, and they use the finite adaptability approach to solve the latter to exact or approximate optimality. The resulting method allows for efficient parallelization, and the authors showcase its effectiveness on route planning and location-transportation problems.