A HIERARCHICAL DEEP REINFORCEMENT LEARNING FOR CAR POOLING WITH AUTONOMOUS VEHICLES

This thesis investigates the use of Hierarchical Deep Reinforcement Learning (DRL) to enhance carpooling strategies for autonomous vehicles. It provides an overview of autonomous mobility and carpooling, examines the carpooling problem. The research presents a hierarchical DRL approach for optimizing pick-up and drop-off strategies, demonstrated through a case study showing improved efficiency and pratical insights.