Software Architecture for a Self-Organizing Logistics Planning System

Abstract

The rise of e-commerce has led to a congested last-mile delivery paradigm. Increasing customer expectations have pushed carriers into a delivery market with diminishing profitability. Furthermore, the current state of last-mile delivery has high societal costs in congestion and environmental impact. To address these challenges, scientists in the logistics field have proposed multimodal transport and collaborative delivery. However, current centralized logistics planning systems are unable to cope with the complexity that these solutions pose. For this reason, Thymo Vlot developed a Self-Organizing Logistics algorithm that leverages decentralization to enable multimodal transport and collaborative delivery. However, the logistics planning system that would utilize this algorithm was left undeveloped, which motivates this thesis project. The first step to develop a software project is to define its architecture. In this thesis project, I analyze Thymo Vlot’s algorithm and develop the software architecture of a logistics planning system that would use it. The main design tool consisted of the selection and application of architectural patterns, which are documented solutions to commonplace problems in software development, drawn from the literature and modern distributed systems. The result of this project is an event-driven microservices architecture, which is highly granular, modifiable, and scalable. These characteristics give the project significant commercial value, as the architecture can be applied to different use cases with different algorithms. The project also leaves behind an architecture with small components, which eases the development cycle of the logistics planning system, thus addressing important managerial challenges. Finally, this project makes significant scientific contributions by developing a software solution that addresses managerial, societal, and environmental challenges of last-mile delivery, thus providing a stepping stone for further research that bridges the gap between the logistics and computer science fields.