Hypergraph Neural Network for Flexible Job Shop Scheduling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current flexible job shop scheduling methods face challenges in handling complex and variable scheduling scenarios, particularly in large-scale systems, due to limitations in scalability and accuracy, with existing reinforcement learning methods struggling to model dynamic states effectively.
Innovation Solution
The proposed method employs a hypergraph-based approach using a hypergraph neural network to extract job features and a policy network to determine actions, combining hypergraph-based and sequential order dependency relationship-based features to improve scheduling accuracy and scalability, while reducing candidate actions using k-nearest neighbors graphs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional reinforcement learning methods are used for flexible job shop scheduling, then the system can handle basic scheduling tasks, but the accuracy and scalability deteriorate when facing large-scale and complex scheduling scenarios
Solution Approach 1:
The patent segments the scheduling system into multiple independent components: state representation module, hypergraph construction module, feature extraction module (with job feature extractor and machine feature extractor), action selection module, and reward calculation module. Each component handles specific aspects of the scheduling problem independently, improving overall accuracy while managing complexity through modular design.
Solution Approach 2:
The patent introduces hypergraph theory as an additional dimensional framework to represent scheduling states. By constructing hypergraphs that capture complex relationships between jobs, machines, and operations beyond traditional graph structures, the system enhances feature extraction capability and scheduling accuracy for large-scale scenarios.
2Adaptability or versatility
If traditional scheduling methods are used, then the system structure remains simple, but the ability to model dynamic states and handle complex scenarios deteriorates
Solution Approach 1:
The patent implements dynamic state modeling through the double-layer attention mechanism that adapts to changing scheduling conditions. The attention weights are dynamically adjusted based on current system state, allowing the model to capture evolving relationships between jobs and machines in real-time scheduling scenarios.
Solution Approach 2:
The patent combines multiple modeling approaches into a composite framework: hypergraph theory for structural representation, attention mechanisms for feature weighting, and reinforcement learning for decision-making. This composite approach enables sophisticated dynamic state modeling while leveraging the strengths of each individual method.
3Measurement precision
If more comprehensive features are extracted to improve scheduling accuracy, then the decision-making quality improves, but the computation time and efficiency deteriorate
Solution Approach 1:
The patent performs preliminary feature extraction by pre-processing job and machine features before the main scheduling decision process. By extracting and organizing relevant features in advance using the hypergraph structure, the system reduces the computational burden during real-time decision-making while maintaining comprehensive feature analysis for accurate scheduling.
4Productivity
If the system handles more job tasks and machines to improve scalability, then the system becomes more versatile, but the existing reinforcement learning methods struggle to maintain effectiveness
Solution Approach 1:
The patent creates a universal scheduling framework that can handle varying numbers of jobs and machines through the hypergraph representation. The hypergraph structure and attention mechanisms are designed to scale with problem size, allowing the same model architecture to effectively solve both small and large-scale scheduling problems without losing reliability.
Data Source
AI summary
A Flexible Job Shop scheduling method includes: obtaining a shop scheduling state including at least one of a sequential order dependency relationship between job tasks being processed, a sequential order dependency relationship between operation steps in each job task of the job tasks, a processing/being processed relationship between the job tasks and machines, or mutual constraint relationships between the machines; representing the shop scheduling state as a state hypergraph; extracting a hypergraph-based job feature from the state hypergraph using a hypergraph neural network; and determining an action configured to change the shop scheduling state, wherein the action is determined according to the hypergraph-based job feature using a policy network.