Please use this identifier to cite or link to this item:
Title: Генетический алгоритм построения потенциальной функции
Other Titles: A genetic algorithm for potential function design.
Authors: Волченко, Е.В.
Keywords: потенциальная функция
генетический алгоритм
фитнесс-функция алгоритма
вероятностная система распознавания
Issue Date: 2007
Publisher: ДонНТУ
Citation: Наукові праці Донецького національного технічного університету. Серія: “Обчислювальна техніка та автоматизація”. Випуск 12(118) — Донецьк: ДонНТУ, 2007
Description: In given clause the technique of construction of a solving rule of likelihood system of recognition is considered by the modified mtthod of potential functions.The method of definition of an opportunity of linear division of classed is described by a cardinal hyperplane. For linearly inseparable classes the genetic algorithm of construction of the potential function,based on likelihood construction of an initial population and a choice of an optimum set of genetic operators is offered. Comparison of results of work of the offered and classical methods of construction of the potential functions, estimating efficiency of use of genetic algorithm with preprocessing for a finding of decision function of trained system of recognition is resulted.
Appears in Collections:Випуск 12 (118)

Files in This Item:
File Description SizeFormat 
133-142.pdf7,13 MBAdobe PDFView/Open

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.