Experimental Study of Convergence and Stability of a Genetic Algorithm Using Different Selection Methods

Genetic algorithms (GA) are indispensable tools in research, enabling the resolution of intricate optimization challenges across diverse domains. Their capacity to thoroughly explore expansive search spaces and bypass local optima sets them apart from conventional approaches. Widely applied in bio informatics for sequence optimization, in engineering for design optimization, and in machine learning for neural network tuning, genetic algorithms demonstrate remarkable versatility in addressing nonlinear, multi-modal problems. This versatility fuels advancements in scientific and engineering research, rendering genetic algorithms vital for innovation and discovery. This article assesses the stability and convergence of a genetic algorithm used to solve a manufacturing problem, where the goal is to maximize performance based on the operation of a set of machines. The objective is to find the most efficient way to operate 10 machines producing various products, utilizing four distinct selection techniques: Fitness Proportional Selection (FPS), Exponential Rank Selection (ERS), Linear Rank Selection (LRS), and random selection. The algorithm's performance will be assessed based on the diversity of solutions generated and the convergence patterns observed for each selection method.