The focus is on the optimal use of the degrees of freedom in the flow of rail traffic in real time, i.e. during operation. A coordinated, optimized selection of the trajectories of the trains on the lines and their stopping times in the stations can reduce the overall electricity demand, which is why a mixed integer program is being developed in the project to determine the most energy-efficient permissible timetable. A prediction model is integrated that uses a Bayesian network to determine forecasts of delays in subway operation.
The optimized timetable model enables the targeted synchronization of departing and arriving trains and thus increases the usable proportion of braking energy fed back into the power grid (so-called recovery rate). In addition, the peak electricity demand can be reduced by avoiding too many simultaneous departures in the grid.
In order to optimize the driving profiles, an iterative process is used that combines the advantages of continuous and discrete optimization methods. Continuous methods enable detailed and realistic modeling while maintaining the efficiency of discrete methods.
The individual components for optimizing the timetable and the travel profiles are brought together in a demonstrator that shows the methods in use. This is intended to illustrate the potential for use in the systems of the application partners.
Division Supply Chain Services at Fraunhofer IIS