Optimizing integrated passenger-freight transport in demand-responsive services: System efficiency and policy insights

This paper examines an Integrated Passenger-Freight Transport (IPFT) approach for Demand-Responsive Transport (DRT) services that serve parcel deliveries alongside passenger trip requests in low-demand or rural mobility contexts while preserving passenger service reliability. The problem distinguishes passenger and freight services through their flow structures, capacity dimensions, service-time components, time-window treatments, and cost representations. It is formulated as a Mixed-Integer Linear Programming (MILP) model that minimizes total system costs, including vehicle and driver operating costs, passenger travel time costs, freight time-window penalties, and emission costs. To solve realistic-size instances, we develop a new hybrid solution algorithm, PSA-GA, combining Population-based Simulated Annealing (PSA) with parent selection and crossover mechanisms from a Genetic Algorithm (GA). Planning and policy implications are assessed through a real-world case study in Kirchheim bei München, Germany, supported by extensive sensitivity analyses on cost structures, demand levels and heterogeneity, temporal demand alignment, travel time uncertainty, and vehicle utilization. Results show that IPFT can reduce total costs by 24 % compared with conventional segregated passenger-freight transport (SPFT) systems. Cost savings increase when freight demand is shifted toward lower-passenger-demand periods, whereas overlapping passenger and freight peaks reduce the benefits of integration. IPFT performance is more strongly linked to overall vehicle utilization than to passenger occupancy alone; although passenger occupancy is slightly lower in the IPFT fleet than in the SPFT passenger-only fleet, the integrated system achieves higher combined utilization through joint passenger-freight routing. Fleet size results further indicate that passenger time value and vehicle operating costs are the most influential planning parameters.