A Data-Driven Status Division Scheme for Automated Container Terminal Production Process based on Graph Information

Automated Container Ports (ACTs) provide crucial linkages in the global container shipping network, the production improvements of ACTs have attracted much attention from academia. The consonant and complete production datasets of ACT can be effortlessly obtained from the Terminal Operation System (TOS), which provides a practical path to describe, analyze, and improve the ACT by data-driven approach in a comprehensive way. Better comprehending and analysis of the data will promote data-driven application of improvement. Therefore, we present a Data-Driven Status Division Scheme for ACT production process. First, a graph representation method is proposed to describe the ACT instantaneous production status. Then, a two-stage feature extraction algorithm is proposed to extract status graph features considering the global information. And then, the production status division scheme is proposed to divide the statuses on the time-continuous production process data. Finally, we illustrate and validate the proposed status division method by the datasets of world's largest ACT and give insights in the status transition of ACT. This study will provide a promising reference for explanatory data-driven decision-making in ACT.