Печатна форма Пояснителен текст Клик - eкран с още данни
СЪЮЗ ПО АВТОМАТИКА И ИНФОРМАТИКА
"Джон Атанасов"

1000 София, ул. “Раковски” 108, тел. 029876169, sai.bg.office@gmail.com, www.sai.bg

Джон Атанасов
JOURNAL "INFORMATION TECHNOLOGIES AND CONTROL"
HOME EDITORS ISSUES TEMPLATE INFORMATION
23/11/2024 PAPERS - ISSUES - ARTICLE

Year 2018 - Issue 4 - Article No 4

POPULATION BASED ALGORITHMS FOR ARTIFICIAL NEURAL NETWORKS TRAINING IN PERFECT INFORMATION GAMES
I. Blagoev, D. Keremedchiev, I. Zankinski
Key Words Perfect information games; artificial neural networks; population based heuristic optimization.
Abstract This research focuses on population based heuristic optimization used for artificial neural networks training. The goalis artificial neural networks to be used as computer opponent in games with perfect information. Artificial neural networks have many applications and one of them is in the game theory. For perfect information games with relatively small board, artificial neural networks can perform well. Training of artificial neural networks with population based heuristics has an advantage that calculations can be done in a distributed computing system, for example a mobile game which communicates with remote server
Article PDF Download article (PDF)