COMPARATIVE ANALYSIS OF THE CONVERGENCE RATE OF PARALLEL ADAPTIVE AND TRADITIONAL GENETIC ALGORITHMS

Chernikov Aleksandr Vladimirovich

The article presents a comparative analysis of two popular optimization methods: the adaptive population algorithm and the traditional genetic algorithm. The main purpose of the study is to evaluate and compare the convergence rate of these two algorithms. The convergence rate is critically important for optimization problems, as it determines the effectiveness of the algorithm in achieving an optimal solution. As part of the study, several test functions were selected, covering a wide range of complexity, from simple non-convex to complex multimodal. This allowed for a comprehensive analysis, evaluating algorithms on a variety of data. Various parameters of the algorithms were considered, including population size, probability of mutation and crossover, as well as selection strategies. The results showed that adaptive population algorithms demonstrate a significant increase in the convergence rate on many test functions, especially in multimodal scenarios. Based on these observations, conclusions are drawn about the applicability and effectiveness of each approach in various areas of optimization. This study represents an important contribution to understanding the dynamics and effectiveness of adaptive and traditional approaches in optimization algorithms, highlighting their strengths and weaknesses in different settings.

Keywords:adaptive population algorithm, genetic algorithm, convergence rate, optimization, multimodal function