Gold Seekers Algorithm: An Innovative Metaheuristic Approach for Global Optimization and Its Application in Image Segmentation

A significant disadvantage of meta-heuristic algorithms is their tendency to get stuck in local optima, which prevents them from finding the global optimal solution and potentially limits their overall effectiveness. To mitigate this drawback, a novel metaheuristic optimization algorithm known as the Gold Seekers Algorithm (GSA) is introduced in this paper to address complex optimization problems. This algorithm mimics the competitive behavior seen among miners hunting for gold. GSA has several key advantages, including fast convergence due to parallel execution, low sensitivity to control parameters due to adaptive coefficients, and high accuracy in global searches thanks to its structured division of search agents into specific roles. The exploitation phase involves searching near the leader to identify valuable areas, while the exploration phase focuses on searching among gold seekers to discover unknown golden locations. To validate GSA's effectiveness, the algorithm was tested on ۲۰ classical benchmark functions of various types and dimensions (۳۰, ۶۰, and ۹۰) and applied to image segmentation to find optimal thresholds. Performance in this practical application was compared to eight well-known heuristic techniques from existing literature.The findings indicate that the suggested algorithm achieves superior solution accuracy and expedited convergence across most benchmark functions, in addition to determining optimal thresholds for image segmentation tasks.

کلیدواژه ها:

Gold Seekers Algorithm ، Metaheuristic Algorithm ، Swarm Intelligence ، Optimization techniques ، Benchmark problems ، Image thresholding