Full Article: PDF
Scientific Object Identifier: http://s-o-i.org/1.1/TAS-04-60-36
DOI: https://dx.doi.org/10.15863/TAS.2018.04.60.36
Language: English
Citation: Korneev AM, Sukhanov AV, Shipulin IA (2018) INVESTIGATION OF THE ACCURACY AND SPEED OF THE ALGORITHMS OF STOCHASTIC OPTIMIZATION FUNCTIONS ON A TWO-DIMENSIONAL SPACE. ISJ Theoretical & Applied Science, 04 (60): 184-189. Soi: http://s-o-i.org/1.1/TAS-04-60-36 Doi: https://dx.doi.org/10.15863/TAS.2018.04.60.36 |
Pages: 184-189
Published: 30.04.2018
Abstract: The purpose of this paper is to analyze the accuracy of calculations and the amount of time spent on finding optimal values for functions of several variables using optimization algorithms based on several methods of stochastic search. To conduct research, the staff of the Department of General Mechanics of the Lipetsk State Technical University created software that implements algorithms for searching for extreme values for functions of several variables. The functional purpose of the software is to find the minimum of a given function, represented as a string of characters. Optimization is performed on a specific and fixed search area, which is a hyperparallelepiped. Each separate program uses its own method of algorithmic optimization. In the development of programs, optimization algorithms based on the Monte Carlo method, an annealing simulation method, an interval analysis method, and a genetic algorithm were used. The results of a computational experiment for three different functions of two variables are presented in the article, a comparative analysis of the closeness of the results to values obtained analytically is carried out. The obtained data allowed us to draw conclusions about the advantages and disadvantages of each of the algorithms. Based on the results of computational experiments, the regularities between the time costs of algorithms and their numerical parameters are determined.
Key words: stochastic optimization, annealing simulation method, Monte Carlo method, genetic algorithm, interval analysis method, convergence rate, extremum.
|
