The Column Generatiion and Traiin Crew Scheduliing

Abstract

Better productivity and efficiency is more and more required in the railway operation. The train crew management is one of the several problems that could be solved using mathematical methods. Crew management is a problem that is well-known in Operations Research. We compare two approaches for solving the train crew scheduling problem. The first approach consists of solving the original problem by single model. The second approach corresponds to the step-by-step column generation. This technique was originally based on Dantzig-Wolfe decomposition. The benchmarks used for comparison of both approaches originate in real problems from railway systems in Slovakia and Hungary.