🐜 ACO with Dynamic Evaporation Rate – Adaptive Metaheuristic for Optimization

This project implements an enhanced Ant Colony Optimization (ACO) algorithm with a dynamic pheromone evaporation rate that adapts over iterations. This adjustment aims to balance exploration and exploitation, improving convergence on both CEC benchmark functions and engineering optimization problems.

---

🚀 Highlights

• 📉 Dynamic Evaporation Rate: The pheromone decay rate increases over time, shifting the search from exploration to exploitation.
• 📈 Performance Evaluation: Validated using CEC Benchmark Suite and real-world engineering test functions.
• 📊 Statistical Testing: Improvements confirmed via Wilcoxon signed-rank tests and visualized with convergence plots.
• 🧰 Tools Used: Python, NumPy, Google Colab, Matplotlib, Git

---

⚙️ Algorithm Overview

• Standard ACO uses a fixed evaporation rate (ρ).

• This version implements:

  ρ(t) = ρ₀ + (ρ_max - ρ₀) * (t / max_iter)
  

  Where:

  • ρ₀ is the initial evaporation rate
  • ρ_max is the maximum allowed decay
  • t is the current iteration

This helps the colony explore broadly in the beginning and narrow down efficiently toward better solutions over time.

---

📊 Results Summary

• Achieved faster and more stable convergence on CEC functions (2014–2020).

• Outperformed baseline ACO in 9/12 benchmark problems (statistically significant at p < 0.05).

• Engineering test cases include:

  • Pressure vessel design
  • Gear train optimization

---

📚 References

• Dorigo, M., & Stützle, T. (2004). Ant Colony Optimization. MIT Press.
• Liang, J.J. et al. (2013–2020). CEC Benchmark Functions
• Wilcoxon Signed-Rank Test – Non-parametric statistical test for pairwise comparisons

---

🤝 Contributing

Contributions, feedback, and pull requests are welcome!

---