Variable Neighborhood Search for Integrated Timetable Based Design of Railway Infrastructure

Abstract

In this paper we deal with the problem of building new or extending an existing railway infrastructure. The goal is to determine a minimum cost infrastructure ful lling the requirements de ned by an integrated timetable and the operation of the railway system. We rst model this planning task as a combinatorial network optimization problem, capturing the essential aspects. We then present a metaheuristic solution method based on general variable neighborhood search that makes use of a dynamic programming procedure for realizing individual connections. Computational experiments indicate that the suggested approach is promising and the analysis of obtained results gives useful hints for future work in this area.