Authors: Max Gath, Thomas Wagner, Otthein Herzog
We present an agent-based approach for solving pickup and delivery problems (PDP) in dynamic environments and show its application by bike couriers in urban districts. To avoid social conflicts, the focus is on the computation of fair allocations of orders to bike couriers who are paid by commission fees. We realize autonomous logistic processes and present a mulitagent system with specially adapted negotiation mechanisms. In addition, we apply the mulitagent-based simulation platform PlaSMA for modeling, simulating, and evaluating different transport scenarios within the transport infrastructure of the city of Bremen, Germany. The customer orders are based on real data provided by a German bike courier company. The results show that the approach indeed improves the fairness of distributions significantly. Finally, we present a new business model for a commercial online dispatching platform enabled by our approach.