A global optimization algorithm inspired in the behavior of selfish herds.

Area of Science:

• Computational Intelligence
• Swarm Intelligence
• Optimization Algorithms

Background:

• Global optimization problems are prevalent in various scientific and engineering domains.
• Existing swarm intelligence algorithms face challenges in balancing exploration and exploitation.
• Simulating natural phenomena offers a promising avenue for developing novel optimization techniques.

Purpose of the Study:

• To introduce a novel swarm optimization algorithm, the Selfish Herd Optimizer (SHO).
• To leverage the selfish herd behavior of animals for solving global optimization problems.
• To enhance the exploration-exploitation balance in optimization without changing population size.

Main Methods:

• SHO simulates predator-prey dynamics using two agent types: selfish herd (prey) and predators.
• Each agent employs unique evolutionary operators based on their role in the simulated interaction.
• The algorithm's performance is evaluated against established methods like PSO, ABC, FA, DE, GA, CSA, DA, MOA, and SCA.
• Standard benchmark functions commonly used in evolutionary algorithm literature are utilized for comparison.

Main Results:

• The Selfish Herd Optimizer (SHO) demonstrated remarkable performance in solving global optimization problems.
• SHO outperformed several well-known evolutionary optimization algorithms in benchmark tests.
• The proposed algorithm effectively balances exploration and exploitation capabilities.

Conclusions:

• The Selfish Herd Optimizer (SHO) is a proficient and robust new algorithm for global optimization.
• SHO offers a valuable alternative to existing methods, particularly in scenarios requiring a refined exploration-exploitation trade-off.
• The simulation of selfish herd behavior provides an effective mechanism for swarm intelligence optimization.