Firefly Mating Algorithm for Continuous Optimization Problems

Area of Science:

• Computational intelligence
• Swarm intelligence algorithms
• Optimization techniques

Background:

• Continuous optimization problems are prevalent in various scientific and engineering domains.
• Existing swarm intelligence algorithms face challenges in global convergence and efficiency.
• Nature-inspired algorithms offer promising solutions for complex optimization tasks.

Purpose of the Study:

• To introduce a novel swarm intelligence algorithm, the firefly mating algorithm (FMA), for continuous optimization problems.
• To enhance the global convergence capability of optimization algorithms through biologically inspired mating behaviors.
• To evaluate the performance of FMA against established algorithms on benchmark functions.

Main Methods:

• The proposed Firefly Mating Algorithm (FMA) utilizes a genetic algorithm core.
• A unique mating pair selection mechanism is inspired by natural firefly mating behaviors, including mutual attraction and multiple mating.
• The algorithm simulates female mating until capacity and male sperm depletion for enhanced population dynamics.

Main Results:

• FMA demonstrated superior performance on 20 benchmark functions compared to 12 other algorithms.
• The proposed algorithm achieved higher success rates in finding global optima.
• FMA required significantly fewer iterations to reach global optima, indicating improved convergence speed.

Conclusions:

• The novel mating strategy in FMA effectively enhances global convergence capabilities for continuous optimization.
• FMA presents a competitive and efficient alternative to existing swarm intelligence algorithms.
• The biologically inspired approach offers a promising direction for future optimization algorithm development.