Districting in last-mile delivery routing: Route creation using shortest Hamiltonian path-based algorithms

• Main Author: Silva, José Miguel Ribeiro e
• Publication Date: 2024
• Format: Master thesis
• Language: eng
• Source: Repositórios Científicos de Acesso Aberto de Portugal (RCAAP)
• Download full: http://hdl.handle.net/10400.22/26252
• topic: Districting; Last-Mile Delivery; Shortest Hamiltonian Path Problem; Routing

Districting can reduce the complexities of delivery problems by segmenting its dimensions while facilitating drivers’ familiarity with their work areas, fostering personal connections with customers, and enhancing satisfaction. This dissertation presents and evaluates multiple heuristics for route creation intending to identify the most efficient method for intra- and inter-district routing. Three types of distances, two TSP-to-SHPP converters, and three types of interdistrict connectors were tested, resulting in a total of 18 distinct variants from the combination of these components. Out of 18 tested variants, the best-performing developed approach used a LinKhernigan-based heuristic, later converting it to a Shortest Hamiltonian Path in each district, creating inter-district connections to a hypothetical medoid in the next district to visit and utilizing asymmetric road distances. The obtained results were satisfactory, and the best components for route creation were identified, allowing for the continuation of the analysis of the model’s adaptability to day-to-day implementation in a further project. The models were developed and tested using real-world data from a parcel delivery company operating in the Porto metropolitan area of Portugal.