Open AccessComparison of methods for start points initializing of a non-parametric optimization algorithm
Author(s) -
A. A. Pavlenko,
В. В. Кукарцев,
В. С. Тынченко,
А. А. Шигина,
Е. А. Чжан,
В. В. Кукарцев
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 - 10.1088/1742-6596/1353/1/012104
Subject(s) - initialization , parametric statistics , algorithm , sequence (biology) , test functions for optimization , mathematical optimization , global optimization , genetic algorithm , computer science , optimization problem , meta optimization , distribution (mathematics) , mathematics , multi swarm optimization , statistics , programming language , mathematical analysis , biology , genetics
The problem of global optimization arises in various fields of science and technology, and several different ways of solving it have been proposed. The results of the study of the effectiveness of the non-parametric global optimization algorithm are presented. A comparative analysis of this algorithm is presented. performance analysis of the algorithm based on the Ackley, Rastrigin, Shekel, Griewank and Rosenbrock function. In addition, studies were carried out for the three initial points of the distribution algorithms: the sequence LPτ, the sequence UDC, the uniform random distribution. thus, the best way to initialize the initial points of the non-parametric optimization algorithm on these test functions was identified. According to the research results, the effective parameters of the genetic algorithm were established.