Open AccessProblem of the selection of genetic algorithm initial configuration
Author(s) -
N. A. Timofeev,
Pavel Peresunko,
S. R. Nekhonoshin,
В. В. Кукарцев,
V. S. Tynchenko,
А. С. Михалев
Publication year - 2019
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1353/1/012113
Subject(s) - selection (genetic algorithm) , set (abstract data type) , genetic algorithm , algorithm , computer science , mathematical optimization , genetic programming , mathematics , machine learning , programming language
The genetic algorithm is one of the most well-known and frequently used global optimization algorithms. The software implementation of the algorithm in many programming languages and the development of various modifications of the selection parameters only encourage authors to analyze the GA parameters and search for their optimal values. Many works are trying to answer the question: what values of input parameters should be used for a greater likelihood of successfully finding the result. However, they only consider a specific set of parameters. This article describes the entire set of GA output parameters and draws up a recommendation for choosing initial values. Favorable sets of parameter values were found. Based on these sets, it is possible to customize them for specific tasks. This study provides an example of using these sets to solve the problem of determining the parameters of a welded beam. The results were obtained corresponding to the best values of similar studies. It is important to note that in solving this problem, an initial set of parameters was already formulated, which facilitated the search for a global minimum.