Reflections on Robust Optimization

Abstract

Various aspect of the robust optimization approach [11] are discussedin the context of scenario based stochastic linear programs. Themain items are the choice of the model parameter, which can be relatedto properties of nonlinearly perturbed linear programs [10] or ofparametric quadratic programs [1], and an extension of the firstresults on the robustness of the optimal value with respect to probabilitiesof the selected scenarios a4d with respect to out-of-sample scenarios,cf. [5].