Testing of Python Models of Parallelized Genetic Algorithms

The paper describes the testing of three models (master slave, fine-grained, and coarse grained) of parallelized genetic algorithms and the comparison of their computational time with each other and with the basic serial model. The analysis of the number of iterations, the load of the main memory an...

Full description

Saved in:
Bibliographic Details
Published in2020 43rd International Conference on Telecommunications and Signal Processing (TSP) pp. 235 - 238
Main Authors Skorpil, Vladislav, Oujezsky, Vaclav, Tuleja, Martin
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.07.2020
Subjects
coarse-grained
Computational modeling
fine-grained
Master-slave
parallelized genetic algorithm
Python
Servers
Sociology
Statistics
Testing
Ubuntu
Workstations
Online AccessGet full text
DOI10.1109/TSP49548.2020.9163475

Cover

More Information
Summary:The paper describes the testing of three models (master slave, fine-grained, and coarse grained) of parallelized genetic algorithms and the comparison of their computational time with each other and with the basic serial model. The analysis of the number of iterations, the load of the main memory and the central processing unit is the subject of other contributions. Corresponding Python modules have been implemented for these models. A test scenario and a test environment were prepared. Testing was realized on a Linux server with the Ubuntu operating system. A RabbitMQ server creating processes by the SCOOP module on the selected workstation was used. Models have been tested by a single-workstation and multi-workstation scenarios. The tested models bring time savings and efficiency improvement compared to the serial model; the fastest was the fine-grained model.
DOI:10.1109/TSP49548.2020.9163475