Search by:
Year of publication
Author name
Paper title
Мультіагентна адаптація гібридного генетичного алгоритму для навчання нейромереж
Full text (PDF)
UDC: 004.89:004.48
Publication Language: Russian
Stuc. intelekt. 2008; 13(4):463-470
Abstract: The agent-oriented method for adaptation of forming and learning of neuronet to learning selection is suggested. Genetic algorithm with real genetic coding is used for adaptation of neuronet structure. Neuronets learning is performing by means of hybrid genetic algorithm with gradient leader relearning. The intellectual agents are used for obtaining of learning parameters. Their knowledge system is based on “chief-inferior”. Knowledge building is performed by means of clusterization. Organization of calculating process allows to perform distributed calculations in heterogeneous local area networks.
Keywords:
References:
View full text (PDF)