In this paper, we propose a new parcel delivery system consisting of a public train and drones. The train, an already existing mobile platform in our neighbourhood, follows its normal predefined route and timetable to transport passengers. In the meanwhile, some parcels to be delivered to some customers and a delivery drone are stored on its roof. The drone can launch from the train, deliver the parcel to a customer, and return to the train. It can also travel with the train and replace its battery on the roof via an automatic battery swap system. As the parcel delivery system cannot manage the movement of the train, the route and the timetable of the train need to be accounted carefully to schedule the deliveries. We formulate an optimization problem to minimize the total delivery time of a given set of parcels, and we propose two algorithms. Though the exact algorithm gives the optimal schedule, its computational complexity is exponential to the number of parcels. To make the proposed system possible to be implemented in practice, a suboptimal algorithm is developed, which is more efficient than the exact algorithm and can achieve close performance with the exact algorithm. Additionally, we propose a simple but effective strategy to deal with the uncertainty associated with the train’s timetable. Moreover, the proposed algorithm for the single drone case is modified to deal with the multiple drone case. The effectiveness of the proposed algorithms is verified via computer simulations and comparison with existing methods. The results show that the presented approach is more cost-efficient than the Truck only scheme and the Truck plus Drone scheme. Moreover, the parcel delivery time can be reduced and the delivery area can be enlarged compared to the scheme using drones only.

在本文中，我们提出了一种由公共火车和无人机组成的新包裹递送系统。这列火车是我们附近已经存在的移动平台，它遵循其正常的预定路线和时间表来运送乘客。同时，一些要交付给某些客户的包裹和一架交付无人机被存储在其屋顶上。无人机可以从火车发射，将包裹交付给客户，然后返回火车。它也可以随火车一起行驶，并通过自动电池更换系统在屋顶更换电池。由于包裹运送系统无法管理火车的行驶，因此需要仔细考虑火车的路线和时间表以安排运送时间。我们制定了一个优化问题，以使给定的一组包裹的总交货时间最小化，并提出了两种算法。尽管精确的算法给出了最佳计划，但其计算复杂度却取决于包裹的数量。为了使所提出的系统能够在实践中实现，开发了次优算法，该次优算法比精确算法更有效，并且可以通过精确算法获得接近的性能。此外，我们提出了一种简单而有效的策略来应对与火车时刻表相关的不确定性。此外，对单无人机案例提出的算法进行了修改，以处理多无人机案例。通过计算机仿真并与现有方法进行比较，验证了所提算法的有效性。结果表明，所提出的方法比“仅卡车”方案和“卡车加无人机”方案更具成本效益。此外，