Egret Swarm Optimization Algorithm: An Evolutionary Computation Approach for Model Free Optimization.

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Bio-inspired Computing

Background:

• Meta-heuristic algorithms are crucial for solving complex optimization problems.
• Existing algorithms like PSO, GA, DE, GWO, and HHO have limitations in certain scenarios.
• Nature-inspired algorithms often provide innovative solutions to computational challenges.

Purpose of the Study:

• To introduce a novel meta-heuristic algorithm, the Egret Swarm Optimization Algorithm (ESOA).
• To leverage the hunting behaviors of Great Egret and Snowy Egret for optimization.
• To evaluate ESOA's performance against established algorithms on benchmark and engineering problems.

Main Methods:

• Development of ESOA based on sit-and-wait, aggressive, and discriminant strategies.
• Implementation of a pseudo gradient estimator for the sit-and-wait strategy.
• Utilizing random wandering and encirclement for exploration in the aggressive strategy.
• Balancing strategies with a discriminant model and incorporating a parallel framework.
• Parameter learning through historical information for adaptability and stability.

Main Results:

• ESOA demonstrated superior effectiveness and robustness compared to PSO, GA, DE, GWO, and HHO.
• ESOA achieved optimal results in all unimodal benchmark functions.
• The algorithm attained high statistical scores, exceeding 9.9 on average and reaching 10.96 and 11.92 on complex functions.

Conclusions:

• ESOA is a highly effective and robust meta-heuristic optimization algorithm.
• The algorithm's unique hunting-inspired strategies contribute to its superior performance.
• ESOA shows significant potential for application in diverse optimization scenarios.