Multi-agent Reinforcement Learning for Collaborative Transportation Management (CTM)

• Published in: Agent-Based Approaches in Economics and Social Complex Systems IX Vol. 15; pp. 123 - 136
• Main Authors: Okdinawati, Liane, Simatupang, Togar M., Sunitiyoso, Yos
• Format: Book Chapter
• Language: English
• Published: Singapore Springer Singapore Pte. Limited 2017; Springer Singapore
• Series: Agent-Based Social Systems
• Subjects: Agent-based modelling; Collaborative Transportation Management (CTM); Economic systems; Knowledge management; Multi-agent system; Reinforcement learning; Research & development management
• ISBN: 9811036616; 9789811036613
• ISSN: 1861-0803
• DOI: 10.1007/978-981-10-3662-0_10

Collaborative Transportation Management (CTM) is a model collaboration in transportation area conducted through information and resources sharing. Planning and implementing CTM not only involve optimization of decisions for all collaborative agents but also involve the influence of different interactions among agents to achieve higher CTM benefits. This paper explores how agent-based modelling is used to model interaction and learning process in CTM in real systems. Agent-based model is used in this paper based on consideration that agent-based model can model the emergent decision patterns and unexpected changes of decision based on the decision-making structure. Model-free reinforcement learning is used to predict the consequences and optimize all agents’ action in CTM.