A customizable metaheuristic framework for solving multi-stage no-wait flowshop scheduling problems, based on real-world academic research.
This project focuses on solving the multi-stage no-wait flexible flowshop scheduling problem (NWFSP), where jobs must pass through several machines in strict sequence, and no job can wait between stages. It's a classical NP-hard scheduling problem, often found in manufacturing, production, and logistics.
Minimize the makespan (total completion time) by assigning job sequences optimally across machines, using metaheuristic techniques such as:
- Particle Swarm Optimization (PSO)
- Genetic Algorithms (GA)
- Tabu Search (TS)
- Modular solver design using the DEAP library
- Simulated benchmark job data
- Plotting of Gantt charts and convergence
- Configurable job instances and machine layouts
MetaFlowScheduler/
├── data/ # Simulated benchmark job data
├── notebooks/ # EDA, experimentation, and visualization
├── results/ # Logs, convergence plots, Gantt charts
├── src/ # Core solver implementations
├── requirements.txt # Python dependencies
└── README.md # Project documentation
- Clone the repo:
git clone https://github.com/mageed-ghaleb/MetaFlowScheduler.git
cd MetaFlowScheduler
- Install dependencies:
pip install -r requirements.txt
- Run a solver:
python src/run_pso_solver.py
- Convergence plots of fitness over iterations
- Gantt chart visualizations of job schedules
Developed by Mageed Ghaleb – Co-Founder of MetaForge | Optimization & AI Specialist
Based on peer-reviewed research in scheduling, metaheuristics, and industrial optimization.
MIT License – Free to use with attribution.