UDC 004.855.5
THE DEVELOPMENT AND STUDY OF HYBRID VERSIONS OF PARTICLE SWARM ALGORITHM BASED ON GRID SEARCH ALGORITHMS
L. A. Demidova, PhD (technical sciences), full professor, RSREU, Ryazan; This email address is being protected from spambots. You need JavaScript enabled to view it.
I. А. Klyueva, post-graduate student, RSREU, Ryazan; This email address is being protected from spambots. You need JavaScript enabled to view it..
The present paper considers the approach to the problem solution of unconstrained optimization based on the hybridization of particle swarm optimization algorithm (PSO-algorithm) and grid search algorithm. The aim of this work is the development of hybrid versions of PSO-algorithm and study of its search characteristics. The paper presents two hybrid versions of basic PSO algorithm, involving the use correspondingly of classical Grid Search (GS) algorithm and Design of Experiment (DOE) algorithm. It is proposed to use canonical PSO-algorithm as base algorithm. The results of experimental studies confirming the application efficiency of the proposed hybrid versions of basic PSO-algorithm in solving optimization problems are represented. Herewith the comparative analysis of main quality indicators of basic PSO-algorithm and its hybrid versions in the problem solution of finding the global optimum of several test functions has been carried out. In addition, the application's expediency of hybrid versions of basic PSO-algorithm in order to reduce the time expenditures for searching of optimum parameters’ of SVM-classifier is shown.
Key words: particle swarm optimization algorithm; grid search algorithm; hybrid algorithm; test function; classification, SVM-classifier, optimization parameters, radial basis kernel function.