DETERMINING THE SIZE OF THE POPULATION OF THE GENETIC ALGORITHM FOR THE PROBLEMS OF DISCRETE OPTIMIZATION IN CAD
A method for determining the size of a population is proposed. The general approach for determining the size of a population follows from the statement that the chromosomes of a population must contain the maximum number of different values that cover most of the search area. The method is based on...
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Belarusian National Technical University
2018-10-01
|
| Series: | Sistemnyj Analiz i Prikladnaâ Informatika |
| Subjects: | |
| Online Access: | https://sapi.bntu.by/jour/article/view/220 |
| Summary: | A method for determining the size of a population is proposed. The general approach for determining the size of a population follows from the statement that the chromosomes of a population must contain the maximum number of different values that cover most of the search area. The method is based on the regression model, which allows you to determine the size of the population, depending on the permissible number of values of the independent variable. The regression model is obtained as a result of processing simulation data in the formation of a population for a single-variable objective function. The problem is solved for a genetic algorithm, where the genotype is represented by a chromosome in binary code, and the phenotype by a decimal integer code of values of independent variables. This allows you to model the formation of a population without reference to specific values of variables. The model was obtained for the power range of the reference sets from 12 to 52, and allows to predict the size of the population beyond the limits of this range. The main area of use of this method is discrete optimization problems with objective functions of several variables, where the ranges of admissible values are finite and have a small dimension. |
|---|---|
| ISSN: | 2309-4923 2414-0481 |