Cost-Benefit Analysis of Investments into Railway Networks with Evolutionary Optimization

Abstract

Abstract The performance of a railway network can be considerably improved if the underlying operational plans, in particular the timetables, are optimized by mathematical methods. For a successful application the optimization tool must be able to reflect multiple conflicting goals as they result from the different viewpoints of the customers (passengers) and the railway company : e.g. short waiting times in stations, high reliability of the timetable vs. small investments into the system, small number of circulating vehicles etc. A result of this multi-criterial approach is not a single optimal timetable but a cost-benefit curve (or surface) which allows to investigate questions as e.g. :”How much money do we have to invest in order to achieve a certain reduction in waiting time for the passengers (or even the integrated time table)?” or”What is the effect to the passengers of a decrease in the rolling stock?” In this paper we present an approach based on modern methods of soft computing. Evolutionary optimization methods are used to find optimal timetables with respect to the criteria ’waiting times’, ’investments into tracks’ and ’no. of vehicles’. The results can be displayed in an interactive graphical user interface that allows to study the solutions in greater detail, e.g. to find out where the investments have to be placed for optimal benefit. Future extensions will cope with more complex scenarios. In particular, we shall include small random delays, as they typically appear during the real operation, into the optimization.