Learning to Evaluate Routes for AGVs in a Port Container Terminal

L. Choi, T. Park, J. Kang, and K.R. Ryu (Korea)


Machine Learning, AGV, and routing


In modern automated container terminals, automated guided vehicle (AGV) systems are considered to be a viable option for the horizontal transportation of containers between stacking yards and quayside cranes. Routing—the process of setting up a travel path for an AGV—is essential for utilizing AGVs efficiently. In this paper, we apply machine learning algorithms to routing in order to improve the efficiency of AGV operation. The proposed method uses a route evaluation model obtained by learning from the AGV travel data, evaluates candidate routes using the evaluation model, and finally selects the best route. The effectiveness of the proposed method was validated by a simulation test. An analysis of the result showed that the routes selected bypassed the congestion areas and were short in length.

Important Links:

Go Back