Auto-encoder Supply Chain Signatures for Efficient Planning Updates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Re-solving supply chain planning problems after minor changes can be inefficient, taking as long as the initial solution, despite few changes and known updates.
Innovation Solution
The implementation of an auto-encoder system that generates supply chain signatures as unique vector space representations, allowing for efficient comparison and updating of supply chain structures by transforming models into digital images and using machine learning to decode these signatures for similarity analysis and solver-less solutions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the supply chain planning problem is re-solved from scratch after minor changes, then the solution remains accurate and complete, but the time required increases significantly
Solution Approach 1:
The patent segments the supply chain planning problem into multiple components: the base problem (unchanged portion) and the changes (modified portion). By dividing the problem this way, the system can leverage the previously solved base problem and only re-solve the changed portions, thereby reducing re-solution time while maintaining solution accuracy through systematic integration of changes.
Solution Approach 2:
The patent performs preliminary actions by maintaining an archive of previously solved supply chain problems and their solutions. When minor changes occur, the system retrieves the pre-computed base solution from the archive rather than starting from scratch, allowing for efficient updates while preserving the accuracy of the original solution through structured modification processes.
2Reliability
If the entire supply chain planning problem is re-solved to ensure accuracy, then the solution quality is maintained, but the computational resources and time are wasted
Solution Approach 1:
The patent extracts the unchanged base problem from the modified supply chain planning problem. By separating the stable components that don't need re-solution from the changed components that require updating, the system maintains solution quality for the entire problem while avoiding redundant computation on unchanged portions, thereby improving planning efficiency.
Solution Approach 2:
The patent applies parameter changes by systematically modifying only the affected portions of the solution when changes occur in the supply chain problem. Rather than re-solving with all parameters, the system identifies and updates only the specific parameters and variables impacted by changes, maintaining solution quality while significantly reducing computational effort and improving productivity.
Data Source
AI summary
A system and method for automated machine learning supply chain planning having a computer with a processor and memory and configured to receive a first supply chain network model having one or more material constraints for operations of a first supply chain network. Embodiments include transforming the first supply chain network model into a digital image, training an auto-encoder model to reduce the dimensionality of an input vector, and locating one or more items in the first supply chain network.