The dynamic Pickup and Delivery Problem with Inside Transfers

In the dynamic pickup and delivery problem with inside transfers transportation requests must be executed by modular autonomous vehicles. A request is defined as the transportation of several passengers from a pickup point to a delivery point. As modular autonomous vehicle can couple while en-route, passengers can use an inside transfer while the vehicles are driving. Outside transfers, in which passengers are transferred between vehicles at a transfer point, are also considered. The objective is to construct vehicle routes that execute as many requests as possible, such that the sum of the total energy consumption of the vehicles and the passenger inconvenience, coming from outside transfers, is minimized. The vehicle routes must satisfy time windows and capacity constraints as well. In this thesis an adaptive large neighbourhood search solution method is formulated to solve this problem. This solution method is included in a rolling horizon approach, to handle the dynamic announcement of requests during the scheduling horizon. The solution method is tested on five instances of 50 requests on five different tree networks. The computational experiments show that inside and outside transfers are used in situations that they are beneficial.