A modelling and algorithmic framework for integrating long-haul transshipment and local transportation

Abstract

Supply chain management comprises many vehicle routing and scheduling problems across different time and geographical scales. In this talk, we discuss a supply chain management problem spanning a large geographical area that integrates customer clustering, transshipment and local transportation. While much research has been performed on each component of our proposed problem, there is currently no established technique for the integration of transshipment with local transportation. This talk will present an iterative, large-neighbourhood search heuristic to find high quality solutions to the integrated transshipment and local transportation problem. We will describe the numerous techniques necessary to diversify the search and solve large-scale supply chain management problems. Our proposed heuristic, based on a multi-armed bandit algorithm, is able to find high-quality solutions for the integrated problem that significantly reduce costs compared to solving each problem sequentially.