AI-Based Container Dwell Time Prediction

Challenge

CTB envisaged to predict container dwell times using artificial intelligence. In the moment of stack-in, import containers lack information on pick-up time and transport means. Yard planning systems calculate the target stacking position based on incomplete information: expected dwell time and ongoing transport are not in the equation. Hence, additional shuffle moves require additional resources, maintenance and energy, leading to inefficiencies.

Tasks Performed

• Developed a machine-learning (ML) model to identify hidden structures, which can be used to predict individual container dwell times and ongoing transport means based on millions of data sets
• Feed those predicted values into the yard planning system

Benefit

Through this approach, HPC assisted the Client in achieving a significant annual reduction of +120,000 shuffle moves with all relevant KPIs reflecting notable enhancements.