Traffic Management

This research topic focuses on proposing and analyzing optimization techniques for various processes at the heart of the railway system, with the aim of increasing efficiency. In particular, most of the research is aimed at exploiting infrastructure capacity and improving service levels.

In this research action, we are tackling, for example, the following problems: saturation of a rail network, operational traffic management, generation of timetables and definition of energy-saving driving profiles. Traditional technologies as well as the latest trends in rail system modernization (e.g., moving block, automatic driving) are considered in the research works.

The techniques proposed to solve these problems are based on operations research and artificial intelligence.

This research is being carried out as part of the development of the RECIFE platform (research on the capacity of railway infrastructures - REcherche sur la Capacité des Infrastructure FErroviaires). This platform was developed in the 1990s and is constantly evolving to meet the needs of the sector.

The RECIFE platform includes a collection of case studies and integrates simulation and optimization tools with analysis and human-machine interface instruments.

Several features distinguish this platform from the majority of tools used in the scientific community.

Microscopic model of infrastructures and train movements

In RECIFE, the modeling of the railway system is based on the microscopic observation of its characteristics and functioning. Virtually all the approaches proposed and the studies carried out consider such modeling.

This implies a considerable data processing effort, and an imperative need of efficiency of the algorithms that must be able to process these large quantities of data.

In return for these challenges, the solutions provided by the algorithms can fully exploit the actual capacity of the system and can be directly deployed.

Integration of Simulation and Optimization

In RECIFE, the algorithms proposed and developed by the traffic management team of the ESTAS laboratory are regularly interfaced with microscopic rail traffic simulators, to study their behavior and performance disregarding the modeling assumptions made. In particular, in these studies, simulation replaces reality, enabling the experimentation to be as close as possible to an actual deployed.

Examples of such experimentation are the studies carried our within the SIGIFRET and the ONTIME projects. In these two projects, the focus is on operational rail traffic management. In SIGIFRET, the optimization algorithm RECIFE-MILP has been connected to the Opentrack simulator. In ONTIME, it has been integrated in closed-loop with the Hermes simulator.

In a recent application, a traffic self-organization approach is integrated in closed-loop with the Opentrack simulator. The latter has also been combined to a microscopic passenger simulator. This application has been developed within the SORTEDMOBILITY project.

Modeling Consistency

In RECIFE, the entirety of the railway transport planning process is considered, taking into account several problems that emerge at different times.

These problems are addressed by proposing closely related models, to ensure that decisions made are consistent at different levels, and to exploit the cross-fertilization of knowledge.

Whenever possible, several problems traditionally treated separately in the literature are solved jointly to improve the overall system performance. For example, the joint resolution of line and timetable planning problems has been successfully studied.