Novas abordagens para problemas de otimização robusta de dois estágios

Abstract

In this thesis, new techniques are proposed to solve the two stages robust optimization problem with uncertain data. First the column and constraint generation method of Zeng and Zhao is speeded by a bundle hybrid strategy, in which the resulting nonlinear problem is solved by an outer aproximation technique. Moreover, a new method is devised. The method combines the Benders-dual with the Zeng and Zhao method. The combination of the method result in a smaller master problem that is easier to solve at each iteration when compared with the column and constraint generation of Zeng and Zhao. Instances of two facility and transportation robust optimization problems were used to assess the performance of the devised solution approaches.