Metaheuristic Algorithms for JSP and FJSP Problems

This repository provides implementations of Genetic Algorithms (GA), Simulated Annealing (SA), and Particle Swarm Optimization (PSO) for solving the Job-shop Scheduling Problem (JSP) and Flexible Job-shop Scheduling Problem (FJSP). These are complex optimization problems frequently encountered in manufacturing and service industries.

---

📖 Introduction

What are JSP and FJSP?

• JSP: A scheduling problem where a set of jobs, each with multiple operations, must be assigned to machines in a specific sequence. The goal is to minimize makespan, reduce idle time, or balance workload.
• FJSP: A variant of JSP that allows each operation to be processed on multiple alternative machines, increasing flexibility but also complexity.

Both JSP and FJSP are NP-hard problems, meaning exact solutions are impractical for large instances. This repository demonstrates the use of metaheuristic algorithms to efficiently find near-optimal solutions.

---

⚙️ Algorithms Implemented

1. Genetic Algorithm (GA): Mimics evolution through selection, crossover, and mutation to explore the solution space.
2. Simulated Annealing (SA): Models the cooling of metals, gradually refining solutions to escape local optima.
3. Particle Swarm Optimization (PSO): Simulates the social behavior of birds to explore the solution space collectively.

These algorithms are ideal for avoiding local minima and exploring large solution spaces efficiently.

---

🚀 Getting Started

1. Installation

Ensure Python is installed, then set up the environment:

pip install -e .
pip install -r requirements.txt

---

2. Data Preparation

You can use the provided example datasets (e.g., mk01.txt) or prepare your own. Custom data should follow this JSON format:

{
    "part1": {
        "process1": [[1, 5], [2, 3]],
        "process2": [[1, 8], [3, 6]]
    },
    "part2": {
        "process1": [[2, 7], [3, 2]]
    }
}

Each process lists the available machines and their respective processing times.

---

3. Running the Algorithms

To run a selected algorithm, execute the following command:

python main.py --solver GA --datapath ./data/mk01.json

Replace GA with SA or PSO depending on the algorithm you want to test. You can also modify hyperparameters within the algorithm scripts inside the methods/ directory.

---

📊 Results

Results will be saved to the results/ folder. Each result includes the following:

• Makespan (total completion time)
• Machine utilization
• Gantt chart visualization (optional: for visualizing schedules)

---

📂 Directory Structure

Metaheuristic-Algorithms-For-JSP-and-FJSP-Problems/
│
├── data/                # Sample datasets
├── methods/             # Algorithm implementations
├── results/             # Output files
├── utils/               # Helper functions
├── main.py              # Entry point for running algorithms
├── requirements.txt     # Python dependencies
└── README.md            # Project documentation

---

🔧 Tuning and Customization

To modify the behavior of each algorithm:

• GA: Adjust mutation and crossover rates.
• SA: Modify the cooling schedule or initial temperature.
• PSO: Tune the inertia weight or cognitive/social parameters.

---

📬 Contact

For questions or feedback, feel free to open an issue or contact the repository owner via GitHub: wyh-neophyte or TreetreeZhang.

---