Structural Stability Assessment for Optimal Order Picking in Box-Stacked Storage Logistics.

Area of Science:

• Logistics and Supply Chain Management
• Artificial Intelligence in Operations Research
• Robotics and Automation

Background:

• Traditional order picking in multi-layer box-stacking storage (BSS) can be inefficient due to the need to remove upper boxes.
• Assessing the structural stability of BSS after box removal is crucial for safe and efficient operations.

Purpose of the Study:

• To develop a time-efficient order picking method for BSS.
• To enable direct picking of target boxes by evaluating structural stability.
• To optimize order picking paths and reduce operational time.

Main Methods:

• Proposed a structural stability assessment (SSA) algorithm for BSS.
• Developed a Convolutional Neural Network (CNN) model trained on simulated and real-world BSS data.
• The SSA algorithm generates complement structure images for CNN input to evaluate stability after target box removal.

Main Results:

• The CNN model achieved 95.1% accuracy on test data and 97% accuracy on real-world data.
• The method enables direct picking of target boxes when structurally feasible, bypassing the need to remove upper boxes.
• Validated the algorithm's effectiveness in real-world BSS logistics environments.

Conclusions:

• The proposed SSA method, powered by a CNN, significantly enhances order picking efficiency in BSS.
• This approach allows for optimal path generation, directly accessing target boxes and reducing handling time.
• The validated algorithm is suitable for practical implementation in BSS logistics for improved performance.