Author(s)

S. Scholz & T. Albrecht

Abstract

A scheduling strategy considering demand variation during the day and within the network is outlined. The optimal headway is calculated taking into account demand and supply simultaneously achieving equilibrium. The train service structure is complicated due to different headways for each hour of the day and various parts of the network. Short train headways shall be provided for sections with high demand whereas sections with lower demand (outskirts) request easily recallable departure times. That is why the use of Genetic Algorithms is proposed to transform such an operating scheme into an optimal train service. They allow to find a timetable with non-conflicting train paths, which is characterized by even headways for all sections of the network. The timetable can also be optimized for minimal fleet size or both goals simultaneously. In order to ease the application of such operating strategies, the proposed algorithms were implemented into a computer-aided planning system called ALFa which is described in this paper. 1 Introduction Demand-driven train operation of fully automated underground railway lines can provide capacity strictly according to the fluctuation of demand. This helps avoid overcapacity during off-peak times and improves the revenue-cost-relation of the operator due to reduced operational efforts. Since this strategy has only been applied to single lines yet [1] the feasibility must be proven for a more extensive network of lines and their temporal and spatial demand distribution. The basic approach of demand-driven operation is characterised by the control of the train headway TS in such a way that transport supply (capacity) is adapted to demand as much as possible. In the early morning and the late evening the longest train headway is chosen short enough so that the passengers do not need to

Keywords