Air-ground multimodal transport planning for joint passenger mobility and parcel delivery: integration of drones, aircraft, and ground vehicles

• We propose an optimization model that jointly schedules heterogeneous eVTOL, drone, and ground-vehicle fleets to meet passenger and parcel demand. • The proposed approach can switch between parcel-only, passenger-only, and combined operations without altering its mathematical formulations. • We customize an Adaptive Large Neighborhood Search algorithm that delivers near-optimal solutions for large instances within practical runtimes. • Experiments on earthquake-response logistics and routine urban services show that integrated air-ground operations cut total cost and delay markedly relative to single-mode baselines, and illuminate trade-offs among vehicle flexibility, speed, and costs. Drones and electric Vertical Takeoff and Landing aircraft (eVTOLs) offer potential to improve transport efficiency and flexibility by circumventing ground transport constraints such as traffic congestion and infrastructure limitations. However, aerial vehicles alone often cannot meet transport demands, necessitating integration with ground-based systems. To address this challenge, this study investigates the Integrated Air-Ground Multimodal Transport Planning Problem (IAG-MTPP) for joint passenger mobility and parcel delivery. The IAG-MTPP integrates fixed-route and flexible-route vehicles, accommodating operational scenarios involving eVTOL aircraft, drones, and ground vehicles. The proposed IAG-MTPP is formulated as a Mixed-Integer Linear Programming (MILP) model that optimizes multimodal operations to minimize transshipment costs, delay penalties, and carbon emissions, while also incorporating capital costs and battery-energy constraints for electric air and ground vehicles. The model is tailored to the complexities of air-ground vehicle routing, considering transfer operations, inter-modal coordination, and routing flexibility. It enables switches between parcel-only, passenger-only, and integrated operations, providing adaptability to evolving transport demands. To solve large-scale instances, we customize an Adaptive Large Neighborhood Search (ALNS) algorithm with insertion, removal, and swap operators, feasibility checks, and an operator-selection scheme. Benchmarking against a commercial exact solver and multiple heuristic algorithms demonstrates the robustness and scalability of the proposed ALNS. The effectiveness and applicability of the proposed model and algorithm are validated through numerical experiments in scenarios including emergency rescue operations in Jiuzhaigou and urban transport in Chengdu, China. The results demonstrate that the integrated air-ground system optimizes multimodal routing while reducing transport costs and improving service coverage compared to alternatives. The proposed ALNS algorithm solves large-scale instances efficiently where commercial exact solvers fail, supporting real-world deployment in large networks and high-demand settings. This study offers insights and guidelines to support efficient resource allocation in the rapidly evolving low-altitude economy.