AI FERODATA Application Enriched with Artificial Intelligence Models to Optimize Freight Transport

Abstract

In freight transportation the planning methods and decision support systems are a crucial point to be considered, yielding interesting research opportunities for the development of optimization. The aim of our project was to research, develop and implement an artificial intelligence (AI) assistant module bringing new AI-based capabilities of optimization and simulation for enterprise-wide operational activities such as management of railway resources and constraints in an efficient and user-friendly manner. The main objective of the current paper concerned the best way to transport freight from given origins to given destinations within time constraints using the railway service provided by the network. This planning problem is faced in four steps (preprocessing, constraints definition, optimization phase and model’s testing). The objective function considers the train operation costs, consumption and time duration. Machine learning algorithms are developed to optimize the objective function according to an enormous number of decision variables and complicated constraints. The platform is tested in a real-world Romania railway network. Future progress of the project will provide models’ testing results and continuous improvement of the algorithms performance.