This PhD thesis (defended in November 2023) from Natasja Sluijk discusses the concept of last-mile delivery in logistics and its associated challenges, the last mile being the most challenging and costly, accounting for over 53% of total shipping costs.
The text discusses advancements in vehicle routing problems (VRPs), which are crucial for various logistics sectors like express delivery, grocery distribution, freight transportation, city logistics, and e-commerce. The focus is on two-echelon VRPs, with an expanded body of literature. This includes mathematical formulations, solution methods, and benchmark datasets for testing new algorithms. Additionally, the text identifies open research areas in this field.
In one of the chapters, the two-echelon vehicle routing problem with stochastic demands is formulated as a chance-constrained stochastic optimization problem. The solution uses column generation, introducing two labeling algorithms for route generation. The chapter emphasizes statistical tests to ensure chance constraints are met and explores methods to simplify these constraints’ verification. Results show the stochastic formulation’s advantages in solution cost and feasibility.
Another chapter introduces a single-echelon stochastic VRP, considering stochastic demand, partial deliveries, and fairness. This involves developing strategies that optimize routing costs and ensure fair distribution of resources among customers, enhancing overall customer satisfaction. This problem is relevant in humanitarian logistics and food rescue. A model is proposed to ensure fair customer distribution with a minimum expected fill rate. A branch-price-and-cut algorithm is developed to solve instances with up to 75 customers, using specific bounding techniques for efficiency. Numerical experiments demonstrate the effectiveness of integrating routing and allocation decisions for fairness, efficiency, and cost-effectiveness compared to sequential approaches.
These chapters highlight significant contributions to optimizing logistics and distribution processes in various sectors, particularly emphasizing two-echelon systems and single-echelon systems focusing on fairness and stochastic demands.
The thesis aims to propose innovative solutions for these extended versions of the vehicle routing problem with stochastic demands, focusing on operational efficiency and sustainability.
Tom Van Woensel 