Kernel-Based Job Scheduling for Adaptive Manufacturing Priority Rules
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing scheduling systems for manufacturing and logistics face challenges in determining optimal job schedules efficiently, especially for large-scale processes, as they often rely on traditional heuristics that may not adapt well to specific environmental constraints and require frequent recomputation.
Innovation Solution
The use of kernel-induced priority functions, which compare jobs to reference jobs using symmetric and positive semi-definite kernel functions, allows for the definition of expressive and adaptive priority rules that can outperform traditional heuristics by incorporating attributes like release dates, processing times, and similarities between jobs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional heuristics are used for job scheduling, then the scheduling process is computationally efficient, but the scheduling performance does not adapt well to specific environmental constraints
Solution Approach 1:
The system employs machine learning models that automatically learn and adapt priority functions from historical scheduling data and environmental constraints without requiring manual intervention. The models self-improve by processing feedback from actual scheduling outcomes, enabling the system to adapt to specific manufacturing or logistics environments autonomously while maintaining high scheduling performance
Solution Approach 2:
The patent transforms fixed traditional heuristics into dynamic priority functions where parameters are continuously optimized based on environmental constraints. The machine learning models adjust weighting factors and priority criteria parameters according to specific environmental conditions, allowing the scheduling system to adapt its behavior while maintaining computational efficiency through learned parameter sets
2Productivity
If complex priority functions are used to improve scheduling performance, then the scheduling becomes more adaptive, but the computational complexity increases
Solution Approach 1:
The system performs computational work in advance by pre-training machine learning models on historical data and environmental constraints. Once trained, the models provide ready-to-use priority functions that can be evaluated quickly during actual scheduling operations. This preliminary training phase separates the computationally intensive learning process from the time-critical scheduling execution, maintaining low computational complexity during operational use
Solution Approach 2:
The patent replaces traditional mechanical scheduling algorithms with machine learning-based predictive models. Instead of using complex rule-based systems that require extensive computational resources to evaluate multiple constraints, the system uses learned priority functions that provide direct predictions of optimal job priorities, significantly reducing computational complexity while maintaining or improving scheduling performance
3Adaptability or versatility
If frequent schedule recomputation is performed to adapt to changing constraints, then the scheduling remains up-to-date, but the computational overhead increases
Solution Approach 1:
The system implements feedback mechanisms where the machine learning models continuously monitor changes in environmental constraints and job characteristics. When changes exceed predefined thresholds, the models update their predictions using new data without requiring complete schedule recomputation. This incremental learning approach allows the system to adapt to changing constraints while minimizing computational overhead by only reprocessing affected portions of the schedule
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
The invention relates to a computer-implemented method (900) of scheduling jobs of a manufacturing or logistics process using a priority function. The priority function is evaluated on multiple to be scheduled jobs to obtain multiple respective priority values. The priority function is defined to invoke a kernel function. The kernel function is defined to compare representations of two respective jobs. Evaluating the priority function on a selected job comprises evaluating the kernel function on representations of the selected job and one or more reference jobs. The schedule for the multiple jobs is determined based on the priority values and is then output to enable the multiple jobs to be carried out according to the schedule.