[FEAT]: Add MLFA-GD: Modified Firefly Algorithm with Gender Difference (Scientific Reports, 2025)

[RdaKA12]

Description

Description

This issue proposes adding MLFA-GD (Modified Firefly Algorithm with Gender Difference),
a recently published variant of the classical Firefly Algorithm.

Note: This algorithm is not the original Firefly Algorithm (FFA) currently implemented in Mealpy.
It introduces new movement rules, population partitioning, and learning mechanisms, making it
algorithmically distinct rather than a simple parameter tuning.

---

Reference Paper (Open Access)

Title: Firefly algorithm with multiple learning ability based on gender difference
Journal: Scientific Reports (Nature), 2025
Link: https://www.nature.com/articles/s41598-025-09523-9

---

Key Differences from Classical Firefly Algorithm

• Population is divided into male and female fireflies
• Different movement strategies are applied based on gender
• Multiple learning mechanisms (centroid guidance, adaptive random walk)
• Improved exploration–exploitation balance
• Demonstrated superior performance over classical FA and several recent optimizers

---

Benchmark and Experimental Setup (Fully Reproducible)

The paper provides explicit experimental settings, enabling near-exact reproduction:

• Benchmark suite: CEC 2017
• Population size: 50
• Dimension: 10
• Maximum iterations: 1000
• Independent runs: 30
• Evaluation metrics: mean and standard deviation

These clear settings allow fair and consistent comparison.

---

Why Add MLFA-GD to Mealpy?

• Open-access, peer-reviewed, recent (2025)
• Clearly documented algorithm and pseudocode
• Explicit benchmark configuration
• Complements existing FA implementation as an advanced variant
• Useful for researchers studying improved swarm intelligence methods

---

Implementation Notes

• Suggested class name: MLFA_GD or Firefly_MLFA_GD
• Can be placed under the Swarm-based category
• MATLAB reference implementation is available in the paper

If maintainers prefer, this can also be categorized as a variant of Firefly Algorithm
rather than a standalone optimizer.

I am willing to submit an implementation and benchmark results upon approval.

Additional Information

No response