Container Loading Management System Using Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing container loading management systems do not consider constraints such as loading balance and are unable to adapt to sequential changes, leading to inefficiencies that depend on the operator's skill level, and require frequent model updates, increasing engineer workload.
Innovation Solution
A container loading management system that includes a container management device, a loading planning device, and a learning device using machine learning to determine optimal loading positions based on real-time data and past records, reducing the need for frequent updates and maintaining accuracy while controlling engineer workload.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a static loading management system is used, then the system structure is simple, but the system cannot respond to sequential changes and loading efficiency depends on operator skill level
Solution Approach 1:
The system transitions from a static management approach to a dynamic one by continuously acquiring current loading state information and sequentially determining loading positions based on changing conditions. The loading position determination unit dynamically adjusts decisions based on real-time container information and loading state, enabling the system to adapt to sequential changes while maintaining manageable complexity through automated decision-making.
2Measurement precision
If the model is frequently reviewed and updated to maintain accuracy, then the model accuracy is maintained, but the workload on engineers becomes large
Solution Approach 1:
The learning device automatically learns and updates the loading position determination model using training data generated from actual loading operations. The system self-updates by acquiring training data from the loading management device and learning device, processing this data to improve the model without requiring manual engineer intervention for each update cycle, thereby maintaining high accuracy while minimizing engineer workload.
Solution Approach 2:
The system implements a feedback mechanism where the loading management device outputs training data based on actual loading operations and determination results. This feedback loop allows the learning device to continuously improve the model by learning from real operational data, maintaining model accuracy through automated feedback-driven updates rather than manual review and adjustment.
3Productivity
If automated loading position determination is implemented, then loading efficiency is improved, but the system does not consider loading balance constraints
Solution Approach 1:
The system incorporates loading balance constraints as preliminary conditions in the loading position determination process. The loading position determination unit considers not only efficiency factors but also constraint satisfaction when determining loading positions. By embedding constraint checking into the determination logic beforehand, the system achieves both high loading efficiency and reliable compliance with loading balance requirements.
Data Source
AI summary
The loading container information input means 71 accepts input of information on the target container. The Inquiring means 72 transmits current loading state and information on the target container to the container loading planning device 80 to inquire about the loading position of the target container. The evaluation means 73 outputs an evaluation value for loading the target container at the received loading position. The output means 74 outputs data including the loading state and information of the target container, the loading position of the target container, and the evaluation value as training data. The learning means 91 learns the model by machine learning using the output training data. The loading position determination means 81 determines the loading position of the target container using the learned model.