A simulation-optimization approach to solve the first and last mile of mass rapid transit via feeder services

This paper focuses on the design of optimal feeder bus routes aimed at serving the commuting demand of suburban areas and increasing the ridership of a mass rapid transit system. The optimization problem presents a multi-objective nature. The transit operator is interested in maximizing the number of users served with the lowest vehicle kilometres travelled (i.e., maximizing profits), whereas passengers seek a high quality of service (i.e., minimizing travel times). An Ant Colony Optimization algorithm is here implemented into an agent-based modelling environment to find the optimal set of routes connecting the service area to multiple transfer stations. The potential demand at a feeder bus stop is estimated according to accessibility indicators, derived from GIS-based demographic data. The model is applied to the case study of Catania (a medium-sized city in Italy) to enhance the accessibility of urban railway stations via public transport. The proposed methodology can be used as a decision-making tool for transport operators and public administrations to understand how to design feeder bus routes to improve the accessibility of public transport.