Robust Algorithms and Price of Robustness in Shunting Problems

In this paper we provide efficient robust algorithms for shunting problems concerning the reordering of train cars over a hump. In particular, we study algorithms able to cope with small disruptions, as temporary and local availability and/or malfunctioning of key resources that can occur and affect planned operations. To this aim, a definition of robust algorithm is provided. Performances of the proposed algorithms are measured by the notion of price of robustness. Various scenarios are considered, and interesting results are presented.