Bi-level optimization for the planning of logistics service networks

Background

In the context of increasing digitalization and advancing globalization, transport service providers are confronted with ever-higher demands in terms of reliability and flexibility because of the processes of change in the production and warehousing of industrial companies. When planning logistics service networks, which are a key success factor for transport service providers, customer behavior has often gone unnoticed until now. For example, customers can choose between standard and express delivery with many service providers. Often, in a hub-and-spoke network typical for logistics service networks, the standard shipment runs via one or two hubs, while the express shipment takes place as direct traffic. Bi-level optimization methods are suitable for this structure. Here, the service provider is the so-called "leader" who first decides on the structure and pricing of his network, while the customer is the "follower" who can then choose between standard and express shipping.

In cooperation with the Chair of Discrete Optimization headed by Prof. Dr. Christoph Buchheim, the Institute of Transport Logistics is working on the DFG-funded project "Bi-level optimization for the Planning of Logistics Service Networks".

Procedure

The basis of the research is a mathematical model that considers the customer's budget as the basis for decision-making. Various optimization methods, such as exact algorithms or heuristics, are to be developed for this model. In the further course of the project, the model will be extended, for example to include further decision-making bases of the customers, vehicle capacities or uncertainty factors. It can be assumed that the optimization procedures will also have to be adapted for this purpose. Finally, the developed procedures will be evaluated with regard to their quality, runtime and practicality.

Targeted results

The aim of the project is to develop practical algorithms for the planning of logistics service networks through methods of bi-level optimization. These algorithms should take into account possible network structures as well as pricing for standard and express shipping.

Contact: Prof. Dr.-Ing. Uwe Clausen , Nele Pommerening

Funding and partners