Automated Bin Packing with Look-Ahead Virtual Sorting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The challenge of efficiently packing mixed-sized boxes in bins using industrial robots is exacerbated by limited online information, leading to increased bin usage and costs, as existing heuristics lack the capability to utilize look-ahead information effectively.
Innovation Solution
A processor-implemented method and system for automated bin placement using virtual sorting and packing, which iteratively receives objects and look-ahead information to perform virtual sorting and packing, determining the correct order and position of objects within bins, adapting standard bin packing algorithms to utilize look-ahead data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If online bin packing is performed with only next upcoming box information, then real-time decision making is enabled, but space utilization becomes suboptimal
Solution Approach 1:
The system performs preliminary sorting and packing operations in advance using available look-ahead information about upcoming boxes. By pre-processing the packing arrangement for boxes that are known to be approaching, the system optimizes space utilization before actual placement occurs, resolving the contradiction between real-time speed and optimal packing.
Solution Approach 2:
A virtual simulation environment acts as an intermediary between the incoming box stream and the physical packing system. The virtual simulation processes look-ahead information and generates optimized packing sequences, which then guide the real-time robotic packing operations, enabling both speed and optimality.
2Ease of operation
If more bins are used to compensate for poor packing efficiency, then packing simplicity is maintained, but costs increase
Solution Approach 1:
The system dynamically changes packing parameters such as box orientation, placement position, and bin selection based on look-ahead information about upcoming boxes. By adjusting these parameters in advance, the system achieves better space utilization with the same number of bins, reducing costs while maintaining automated operation simplicity.
3Quantity of substance
If look-ahead information is utilized for virtual sorting and packing, then space utilization improves, but computational complexity increases
Solution Approach 1:
The computational problem is segmented into smaller sub-problems by processing boxes in batches based on look-ahead windows rather than attempting to optimize all boxes simultaneously. This segmentation reduces computational complexity while still capturing the benefits of look-ahead information for improved space utilization.
Solution Approach 2:
A virtual copy or simulation model of the packing process is created and processed in advance using look-ahead information. This virtual model handles the computational complexity, while the physical system only needs to execute pre-determined packing decisions, reducing real-time computational requirements.
Data Source
Figure 1
Figure 2
Figure 3A
AI summary
Online 3-dimensional bin packing problem (O3D-BPP) is getting renewed prominence due to the industrial automation brought by Industry 4.0. However, due to limited attention in the past and its challenging nature, a good approximate technique is in scarcity as compared to ID or 2D problems. Present disclosure provides system and method that considers real-time O3D-BPP of cuboidal boxes with partial information (look-ahead) in an automated robotic sorting center. System presents two rolling-horizon mixed-integer linear programming (MILP) cum-heuristic based algorithms wherein a framework is provided that adapts and improves performance of BP heuristics by utilizing information in an online setting with look-ahead.