Improvement of freight consolidation through a data mining-based methodology

Freight consolidation is a complex logistics practice supported by a broad spectrum of strategies and methods to improve supply chain cost-effectiveness. It consists of grouping products in a single batch to reduce distribution costs. Literature review revealed that operational research (OR) is typically used for freight consolidation, and their inputs are often aggregated over time. While necessary to accommodate computationally expensive OR algorithms, such data simplifications are responsible for losing valuable data patterns. Our contribution is a novel data mining methodology that uses association rules to leverage data patterns in the context of intermittent demand. Our approach is compared to a typical operational research approach from a literature case study. Simple to implement, our methodology gives good results and can flexibly accommodate and exploit data patterns while being able to scale to a much larger amount of data, making it a more suitable approach for the big data world.