New Modelling Approaches for Location and Routing Problems Towards Sustainable Logistic Systems

Sammanfattning: Transportations are paramount for a well-functioning society and necessary to secure essential products and maintain our standard of living. These are operations and activities related to distributing packages and goods and providing services by society such as waste collection, postal services, health care, or emergency response. Transports thus affect all of us, companies, the public sector, and individuals, in our daily lives. However, the transportation sector is one of the most polluting sectors, and in recent years the number of parcels delivered has increased significantly. Based on the importance of transportation and logistics services and their impact on the economy, environment, and people’s lives, it is in everyone’s interest that these transports are as efficient as possible.However, the efficiency and design of distribution routes are affected by the underlying infrastructure of the supply chain. Furthermore, as such infrastructure is associated with substantial capital investments, it is equally important to consider the network design and, thus, the location of various facilities such as plants, warehouses, or depots. The field of supply chain network design encompasses two major and well-studied problem classes; the vehicle routing problem aiming to find optimal routes to serve a predetermined set of customers by a fleet of vehicles from a central facility, and the facility location problem aiming to find the optimal location for various facilities. Moreover, these problems are interrelated as the facilities’ location and their respective capacity affect the demand fulfillment and, therefore, the vehicle routing. Such problems are often solved with optimization techniques within Operations Research, a branch of Applied Mathematics concerned with the mathematical modeling and algorithmic solution of decision-making problems. Additionally, changes in the conditions of supply chains have directed the attention of researchers over the last decades to extend and introduce new variants of network design models to cope with real-life characteristics of the transportation system. Incorporating such aspects increases the complexity of operational constraints, making the models harder to solve to optimality.This thesis aims to contribute to the advancement of the field of supply chain network design. Specifically, there are two aims [ I ]  to explore and identify opportunities and needs in modeling and solution methods that cope with real-life aspects and increased complexity to location, distribution, and transportation systems, and [ II ] to propose new modeling approaches and solution methods that cope with such opportunities and needs.The contributions are summarized based on three appended papers. Paper A discusses the difficulties in estimating the demand for a reverse supply chain for a newly introduced product and dealing with such uncertainties to make location decisions. A modeling approach is presented, where each time period for each demand scenario is optimized separately but influenced by the optimization of the previous time periods. Paper B is a literature review covering multi-objective location-routing problems; these are strategic models aiming to determine the location of facilities considering aspects of tour planning and multi-stop routes. An annotated review is presented based on the application area of the various models, and an analysis of objectives and solution approaches used. Paper C introduces the Hierarchical Multi-Switch Multi-Echelon VRP, which is a new variant of the vehicle routing problem and a real-life problem originating from the policies of a Nordic distribution company. A mixed-integer formulation of the problem is proposed, and its relations to other previously stated VRP variants are analyzed and discussed.