Container Terminal Liner Berthing Time Prediction with Computational Logistics and Deep Learning

The quayside running conditions play a key role in container terminal logistics systems, and the terminal liner berthing time (LBT) is the central index of quayside service efficiency that is also the important evidence and guidance to task scheduling and resource allocation at container terminals. The computational logistics and deep learning are combined to discuss the prediction of LBT by the generalization, unification and integration of the essence and connotation of computation. It is supposed to integrate the deep neural networks learning computation and logistics generalized computation for container terminals (LGC-CT) cross the boundaries between information space and physical world. A deep learning model is designed and executed to predict and evaluate LBT at a typical container terminal in China based on its LBT data for the past four years, which is also intended to lay a good foundation for the configuration, deployment and execution of LGC-CT. The deep neural networks are designed and implemented by the fusion of long short-term memory network, gated recurrent unit one, Gaussian noise one and dense one with TensorFlow 2.3, which demonstrates the feasibility and credibility of the proposed compound computing architecture and paradigms preliminarily.