MACHINE LEARNING AND EXPLAINABLE AI FOR SCHEDULING IN DELIVERY LOGISTICS

This dissertation explores the integration of machine learning and XAI to address scheduling and dispatching challenges in three critical logistics environments: UAV-based delivery systems, OHT in semiconductor manufacturing, and last-mile delivery operations, exemplified by the Amazon Last Mile Challenge. Each of these areas presents unique logistical hurdles, from managing UAV payload limitations and battery life to coordinating hundreds of vehicles in OHT systems and balancing drivers’ preferences in last-mile delivery. By investigating these diverse contexts, this work aims to develop and validate novel scheduling solutions that leverage the strengths of machine learning while ensuring explainability through XAI.