Abstract
The area of population-based meta-heuristics has been researched extensively in recent years. The focus of this research has been on finding improvements and variations to existing algorithms while the inner details, that are treated as a black box, remain poorly understood. The purpose of this paper is to uncover the detailed behavior of Variable Mesh Optimization (VMO), a population-based meta-heuristic, and describe the patterns that drive the algorithm in finding new optima. Our results suggest that, in VMO, the improvement of the best solution is strongly correlated with its adaptive clearing mechanism. It is observed that each relaxation of the threshold that is used by the mechanism, is likely to increase the accuracy of the final solution. These findings suggest that future research, aiming to improve algorithm accuracy, could focus on improving the adaptive clearing mechanism in order to increase the likelihood of creating superior algorithms.