CNC Job Scheduling Using Genetic Algorithm Evaluation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional scheduling methods for computer numerical control machines are rule-based and lack clear evaluation criteria, making it difficult to produce optimized schedules that satisfy complex constraints in production environments where thousands of product types are fabricated, leading to inefficient job distribution.
Innovation Solution
A method and apparatus using machine learning, specifically a genetic algorithm, to generate and evaluate multiple schedules based on performance indices such as deadline adherence, delay times, and end times, iteratively improving schedules until a target evaluation index is reached, ensuring optimized job assignment while satisfying production constraints.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If rule-based scheduling methods are used, then scheduling can be performed according to simple conditions, but clear evaluation criteria are lacking and performance evaluation becomes ambiguous
Solution Approach 1:
The patent transforms the scheduling system from rule-based to evaluation-index-based by changing the fundamental parameters used for scheduling decisions. Multiple quantitative evaluation indices (deadline adherence rate, average delay time, machine utilization rate) replace subjective rule-based criteria, enabling precise measurement and comparison of different scheduling schemes.
Solution Approach 2:
The patent replaces the mechanical rule-based scheduling system with a machine learning-based intelligent system. The genetic algorithm and evaluation index system substitute traditional manual or rule-based scheduling mechanisms, enabling automated optimization without relying on predefined rules that lack clear evaluation criteria.
2Speed
If conventional rule-based scheduling is used, then scheduling can be performed quickly, but optimized schedules satisfying complex production constraints cannot be generated
Solution Approach 1:
The patent introduces dynamic optimization through iterative genetic algorithm operations. Instead of static rule-based assignments, the system dynamically generates and evolves scheduling schemes across multiple generations, adapting to complex constraints and continuously improving solution quality through selection, crossover, and mutation operations.
Solution Approach 2:
The patent implements feedback mechanisms through the evaluation index system. Each scheduling scheme is evaluated against multiple quantitative criteria, and this feedback drives the genetic algorithm's selection process. The system uses evaluation results to guide subsequent generations toward better solutions, creating a closed-loop optimization process that balances speed and quality.
3Manufacturing precision
If multiple evaluation indices are calculated for schedule optimization, then schedule quality improves, but calculation complexity and time consumption increase
Solution Approach 1:
The patent segments the complex evaluation process into distinct, modular indices: deadline adherence rate, average delay time, and machine utilization rate. Each index independently evaluates a specific aspect of schedule quality, making the complex evaluation system manageable and interpretable while maintaining comprehensive assessment capability.
Solution Approach 2:
The patent creates a universal evaluation framework that can assess multiple scheduling schemes simultaneously using the same set of indices. This multi-functional evaluation system works across different production scenarios and constraint types, providing consistent quality measurement without requiring separate complex evaluation systems for each case.
4Productivity
If genetic algorithm operations are performed repeatedly to generate new schedules, then schedule optimization improves, but computational time increases
Solution Approach 1:
The patent applies partial action by performing genetic algorithm operations iteratively but with controlled generations. Rather than exhaustively searching all possible schedules, the system performs a sufficient number of iterations to achieve acceptable optimization, balancing computational effort with solution quality through practical convergence criteria.
Data Source
AI summary
The method includes collecting a schedule job list from a database, generating a plurality of schedules for a schedule job to be processed with respect to the schedule job list, calculating an evaluation index for the plurality of generated schedules, determining whether the calculated evaluation index for the plurality of schedules has reached a target evaluation index, selecting a schedule corresponding to two evaluation indices when the calculated evaluation index does not reach the target evaluation index and generating two new schedules using a genetic algorithm, and setting a selection probability so that a schedule having the highest evaluation index is selected and returning the selection probability to a user when the calculated evaluation index reaches the target evaluation index.