UDC 004.272.2
ACCELERATION OF GENETIC ALGORITHM FOR TRANSPORTATION PROBLEM SOLUTION BY MEANS OF MULTITHREADING PROGRAMMING
S. V. Skvortsov, Dr. Sc. (Tech.), full professor, RSREU, Ryazan, Russia;
orcid.org/0000-0001-9495-4953, e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
M. S. Dyakov, master student, RSREU, Ryazan, Russia;
orcid.org/0000-0001-7809-0294, e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
The application of a genetic algorithm for solving a transportation problem according to cost criterion is studied. The aim is to organize parallel processes in the implementation of basic genetic operators, aimed at accelerating the operation of genetic algorithm. Possible approaches to the implementation of multithreaded applications for solving a transportation problem based on a genetic algorithm using modern programming technologies are presented. A computational experiment was performed to evaluate the acceleration of a genetic algorithm when processing large populations using CUDA and OpenMP multithreading programming technologies. The features of Linux operating system architecture related to data buffering when accessing computer memory are investigated, the use of which also makes it possible to increase the efficiency of genetic algorithm when processing large populations. Experimental estimates of a multithreading application acceleration based on OpenMP technology achieved by programmatically disabling data buffering are obtained
Key words: : transportation problem, genetic algorithm, selection, crossover, mutation, sequential algorithm, multithreading, graphics processing unit