Fuzzy C-means clustering based vertical container stacking in container terminals.

Area of Science:

• Operations Research
• Logistics Management
• Artificial Intelligence

Background:

• Container terminals face inefficiencies due to reshuffling, which involves unnecessary container movements during retrieval.
• Traditional stacking strategies often fail to account for container weight variations, leading to operational bottlenecks.

Purpose of the Study:

• To propose a novel fuzzy clustering-based vertical stacking strategy (FVSS) for container terminals.
• To enhance operational efficiency and reduce weight variance in container yard operations.
• To address the challenge of reshuffling by dynamically adapting stacking to container weight.

Main Methods:

• Implemented Fuzzy C-means (FCM) clustering to classify containers into weight classes based on historical data.
• Developed a vertical stacking strategy that proportionally allocates stacking spaces based on cluster sizes.
• Assigned a weight reference value to each stack to guide real-time container stacking decisions.

Main Results:

• FVSS achieved up to a 78% reduction in weight variance compared to existing methods (HSS, RSS, GMM).
• The strategy demonstrated stable performance even under uncertain container weight conditions.
• Numerical experiments using real-world terminal data validated the effectiveness of FVSS.

Conclusions:

• FVSS offers a robust and computationally efficient solution for container yard operations.
• Integrating fuzzy optimization into stacking strategies significantly improves weight management and operational efficiency.
• The proposed method provides a flexible vertical stacking approach that adapts to weight similarities and uncertainties.