Agent learning in fully observable, continuous and real-time game environments


Tezin Türü: Yüksek Lisans

Tezin Yürütüldüğü Kurum: Orta Doğu Teknik Üniversitesi, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü, Türkiye

Tezin Onay Tarihi: 2016

Öğrenci: ÖMER BAYKAL

Danışman: FERDA NUR ALPASLAN

Özet:

Game industry has become one of the sectors that commonly use artificial intelli- gence. Today, most of the game environments need and include artificial intelligence agents to offer more challenging and entertaining experience. Development processes and the quality of artificial intelligence agents are the most important concerns in this area. Since it becomes harder to develop good agents as games become more com- plex, machine learning methods have started to be used in some notable games to shorten this development process and to improve the quality of agents. Popularity of machine learning applications in game environments has increased in last decades. Supervised learning is a machine learning method which can be applied to develop artificial intelligence agents that play a game like human players by imitating them. The imitating agents can either play the role of opponents or play on behalf of the real players when they are absent. The purpose of this study is to develop imitating agents for one of the world’s most played online game; HaxBall. The developed agents can mimic the real HaxBall players. HaxBall is a two dimensional football game with fully observable, continuous, and real time game environment.