A product-driven system with an evolutionary algorithm to increase flexibility in planning a job shop.

Area of Science:

• Operations Research
• Artificial Intelligence
• Manufacturing Systems

Background:

• Traditional job shop scheduling models are computationally intensive and impractical for real-time implementation, especially for large-scale problems.
• Decentralized approaches using real-time product flow information offer a dynamic alternative for makespan minimization.
• The computational performance of decentralized systems like holonic and multiagent systems for real-time job shop control remains unclear.

Purpose of the Study:

• To present a product-driven job shop system model incorporating an evolutionary algorithm to minimize makespan.
• To evaluate the computational performance and solution quality of this system across different problem scales.
• To determine the feasibility of embedding such a system in a real-time production control process.

Main Methods:

• Development of a product-driven job shop system model.
• Integration of an evolutionary algorithm for makespan minimization.
• Simulation of the model using a multiagent system to compare performance against classical models for 102 job shop problem instances (small, medium, large scale).

Main Results:

• The product-driven system consistently produces near-optimal solutions within short timeframes.
• System performance improves as the scale of the job shop problem increases.
• Observed computational performance indicates suitability for real-time control applications.

Conclusions:

• The proposed product-driven job shop system effectively minimizes makespan using an evolutionary algorithm within a multiagent system framework.
• The system demonstrates scalability and efficiency, outperforming classical models for larger problem instances.
• The findings support the practical implementation of this decentralized approach for real-time job shop scheduling and control.