Optimizing the service area and trip selection of an electric automated taxi system used for the last mile of train trips

• Two integer programming models are proposed to optimize automated taxi systems.
• Careful zoning and trip selection of automated taxis increases the profit.
• Having electric automated taxis constrains in accepting more requests.
• Using automated taxis can decrease the relocation costs and improve the profit.

We propose two integer programming models for optimizing an automated taxi (AT) system for last mile of train trips. Model S1: trip reservations are accepted or rejected by the operator according to the profit maximization; model S2: any reservation on a selected zone by the model must be satisfied. Models were applied to a case-study. Results indicate that fleet size influences the profitability of the taxi system: a fleet of 40 ATs is optimal in S1 and 60 ATs in S2. Having electric ATs constrains the system for small fleets because ATs will not have time for charging.