Designing a bus routing model for feeder lines connected to the express bus network in Wallonia

Details

Supervisors

Defourny, Pierre ; Delecluse, Augustin

Faculty

Faculté des bioingénieurs

Degree label

Master [120] : bioingénieur en sciences et technologies de l'environnement, à finalité spécialisée

Abstract

enThis thesis develops a reproducible method for generating feeder bus routes connected to existing express bus lines in middle-density areas. It addresses the difficulty of improving public transport accessibility where demand is spatially dispersed and where conventional high-frequency services are difficult to operate efficiently. The objective is to explore whether a score-based route-generation model can identify relevant feeder corridors under operational constraints. The method combines spatialised user estimation, destination attractiveness and route optimisation. Current and potential bus users are estimated from Wallonia-scale mobility survey data and applied to a population grid. Local stops are scored according to the estimated users located within their walking catchments, while express stops are scored according to the attractiveness of the express lines they provide access to, using points of interest as destination indicators. These scores are then integrated into an OR-Tools route-generation model, which selects optional stops by balancing travel-time costs with visit rewards. Experimental results show that the model can generate plausible feeder routes reflecting local demand, express-network structure and imposed routing constraints. Selected routes are then assembled into a proposed feeder network which, with four distinct lines, connects approximately 50% of the identified users to between three and six express lines. Validation through comparison with existing TEC lines suggests that the model can reproduce realistic route structures when appropriate constraints are imposed, while also replacing some low-scoring stops with higher-scoring alternatives. The proposed framework should be interpreted as an exploratory planning tool rather than a complete operational network design. It suggests improved spatial access to the express network but does not directly measure modal shift or competitiveness with the private car. It does not include frequency setting, timetable coordination, transfer waiting times or passenger assignment, but it provides a structured basis for identifying promising feeder corridors and supporting further operational analysis.