Parallel-SwarmOptimization-Project

Parallel-SwarmOptimization-Project is an optimization project based on the PSO (Particle Swarm Optimization) algorithm, developed to evaluate and compare different parallelization strategies in Python. It was created as part of a practical assignment for the Parallel Programming course.

🧠 Objective

The goal of the project is to compare the performance of various parallel execution models when running multiple parameter combinations of the PSO algorithm across several benchmark functions (Sphere, Rastrigin, Ackley).

The following PSO variants were implemented:

• app.py: sequential version.
• app_threading.py: multithreaded version using ThreadPoolExecutor.
• app_multiprocessing.py: multiprocess version using multiprocessing.
• app_asyncio.py: asynchronous version using asyncio.
• app_multi_asyncio.py: combined multiprocessing + asyncio version.

Each script generates its own CSV output and records its total execution time.

---

⚙️ Project Structure

pso_project/
│
├── src/
│ ├── app.py
│ ├── app_threading.py
│ ├── app_multiprocessing.py
│ ├── app_asyncio.py
│ ├── app_multi_asyncio.py
│ ├── PSO/
│ │ ├── pso_algorithm.py
│ │ └── particle.py
│ └── func/
│ └── objective_functions.py
│
├── analysis/
│ ├── results.csv
│ ├── results_threading.csv
│ ├── results_multiprocessing.csv
│ ├── results_asyncio.csv
│ ├── results_multi_async.csv
│ └── notebook.ipynb
│
├── requirements.txt
└── README.md

---

📦 Installation

This project requires Python 3.9 or later.

Install the dependencies with:

pip install -r requirements.txt

---

📊 Analysis

The notebook notebook.ipynb includes:

• Execution of all algorithm variants from Jupyter.

• A visual comparison of execution times.

• Calculation of Speedup for each version relative to the sequential baseline.

• A brief discussion on parallelization decisions, the GIL, and the impact of using different strategies in CPU-bound scenarios.

---

📝 Notes

• Execution times may vary slightly depending on the system load.

• In this project, the most efficient implementation was the Multi + Asyncio version, which combines multiprocessing with asynchronous task execution.

---

▶️ How to Run

To run any of the PSO implementations manually, open a terminal and navigate to the src directory. Then, execute the desired script using Python. For example:

cd src
python app.py

You can replace app.py with any of the following:

app_threading.py

app_multiprocessing.py

app_asyncio.py

app_multi_asyncio.py

Each script will run the optimization process, print progress to the console, and generate a CSV file with the results in the analysis folder.

• Note: Running all combinations may take a few minutes depending on your system.

---