A new population-based metaheuristic optimization method: Shooting algorithm

In this thesis, a novel metaheuristic optimization algorithm, inspired by the rifle shooting phenomenon and named the Shooting Algorithm, is introduced and its performance is systematically evaluated in comparison with widely used optimization algorithms in the literature. The Shooting Algorithm conceptualizes the search space as a target board, aiming to identify the optimal solution by iteratively narrowing the spread of "shots" toward the target. The algorithm begins with an initial population generated through random shooting, identifies the best candidate solutions, and subsequently refines the search by generating new shots centered around these best solutions. The radius and center of the new shots are determined based on their Euclidean distances, and the shot radius progressively decreases with each iteration, honing in on the optimal solution. To assess the performance of the Shooting Algorithm, a comprehensive benchmark using 16 test functions commonly employed in the literature was conducted. The algorithm's results were compared against those of Genetic Algorithm (GA), Particle Swarm Optimization (PSO), Bee Colony Algorithm (BCA), and Whale Optimization Algorithm (WOA). Performance metrics, including convergence behavior, solution accuracy, and consistency, were analyzed. The results demonstrated that the Shooting Algorithm achieved low error rates and standard deviations on unimodal functions and produced competitive outcomes on multimodal functions. Non-parametric statistical analyses further indicated that the Shooting Algorithm significantly outperformed other algorithms, displaying superior accuracy and consistency across numerous test functions. In conclusion, the Shooting Algorithm is evaluated as a reliable and robust optimization method, particularly excelling in problems requiring high precision. Its performance can potentially be enhanced through the incorporation of dynamic and adaptive mechanisms. The findings suggest that the Shooting Algorithm is a versatile tool with applications across various domains, representing a noteworthy contribution to the field of metaheuristic optimization.