Conference Agenda

The increasing volume of personal motorized vehicles (PMVs) in cities has become a serious issue, leading to congestion, noise, and air pollution. To mitigate these negative externalities, new forms of mobility such as ridepooling promise to redefine urban road traffic by offering mobility as a service. Assuming that this may become a predominant mode of individual transport, urban traffic could potentially become more manageable through centralized routing decisions.

The presented heuristic is built upon this assumption and detects bottlenecks based on traffic simulation. For each bottleneck, the heuristic systematically proposes efficient rerouting options to alleviate traffic congestion, taking into account the severity of congestion and the requirements and priorities of each vehicle. The traffic simulation employed is based on a realistic default traffic scenario for a segment of the city map of Darmstadt, which serves as the initial solution.

The evaluation of the results is based on the resulting utilization of infrastructure and the "travel time quotient" ("Beförderungszeitquotient"), comparing the actual travel times with the minimum travel times that would occur without any congestion. Trade-offs between both parameters are analyzed, and the results are systematically compared to the default solution.

During a day, traffic demand greatly varies; for example, it is expected more traffic in early morning and late afternoon due to people commuting to and from work. As congestion builds up, commuters become more irritated and more susceptible to mistakes. Therefore, searching for improved solutions is fundamental, as one or two extra seconds of green light phase can greatly alternate the traffic situation.

An important indicator for evaluation of traffic is called Vehicle’s Delay Time (VDT), which is defined as the extra time that vehicles need to complete its journey due to congestion or red traffic lights in comparison to a clear way journey. As part of the HafenPlanZen project, a project to create a strategic port planning, our objective is to use a simulation model of a port area of the city of Hamburg in Germany to optimize the phase configuration of traffic light signals in an intersection and to minimize VDT. A simulated annealing optimization algorithm was chosen, because it has an easy implementation and it has been shown to be successful with many problems.

The simulation model is already completed and the focus now is on implementing the optimization algorithm. In order to converge to a solution faster, we plan on using the Downhill Simplex Method developed by Nelder and Mead in 1965. We will present the traffic scenario implemented today and the optimized solution to show the improvement.

Efficient light rail turnaround planning and assignment to overnight parking spaces are essential for optimizing urban public transport systems. This study presents a software-based light rail scheduling tool developed for DSW21, a communal transport company in Dortmund, Germany, to automate a previously manual process.

The challenge of this task lies in meeting the timetable and simultaneously scheduling light rails to accessible parking spaces. Furthermore, the parking areas have unique constraints that restrict the scheduling decision, such as the number of parking spaces available or the consideration that a light rail can only leave if another does not block the departure.

Our tool takes timetable data as input and generates schedules using a First Come, First Served heuristic with constructively assigning light rails to accessible parking spaces. The tool reduces the need for manual schedule readjustments, consistently achieving placement rates of 95% to 100% for routine timetables. Moreover, we achieved user-friendly functionality and respected company standards and design conventions by involving users in the development process. The tool incorporates editable user files to facilitate practical utility and ensure adaptability for future rail modifications. Additionally, our tool tracks light rail maintenance cycles to determine whether the maintenance requirements are met and identifies light rails that must be driven to the depot outside the timetable.

Moving forward, we aim to improve the algorithm to increase placement rates for high-volume timetables and optimize scheduling within maintenance constraints to reduce the number of manual schedule adjustments ex-post, thus continuously improving the company’s light rail operations.

The safety of a street for school children can be modeled by using ordinal costs, i.e., ordered categories like safe, medium and unsafe. Ordinal costs model the quality of objects whenever numerical values are not appropriate, i.e., it is impossible to quantify how much better a safe street is compared to a medium safe street.

In this talk, we present a solution method for solving ordinal shortest path problems by a transformation into associated multiobjective shortest path problems. This allows us to compute a set of efficient solutions with different trade-offs between path length and safety. The choice of a route depends then on the individual preference of the decision maker. However, the safety of a school path is highly depending on the crossing points and not only on the longitudinal traffic. We discuss the difficulties of integrating safety of crossing points and streets in one optimization model and present different modeling ideas.