Sustav za igranje računalnih igara temeljen na dubokom podržanom učenju

master's thesis

• Title (english): Computer Game Playing System Based on Deep Reinforcement Learning
• Author: Filip Štetić
• Mentor: Tomislav Hrkać (mentor)
• Committee member: Tomislav Hrkać (predsjednik povjerenstva)
• Committee member: Zoran Kalafatić (član povjerenstva)
• Committee member: Leonardo Jelenković (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2023-07-10, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Podržano učenje raste u popularnosti razvojem novih metoda i sve većom primje-
nom. Samo područje nije novo i razvoj je započeo mnogo prije nego je ušlo u školske
kurikulume. Igrica Pong popularno je okruženje za testiranje i evaluiranje algoritama
podržanog učenja. U ovom radu daje se pregled povijesti podržanog učenja i objašnja-
vaju se mnogi od temeljnih algoritama zaslužnih za razvoj drugih, danas korištenijih,
algoritama. Daje se pregled popularnih algoritama u dubokom podržanom učenju. Na
kraju se par dubokih algoritama, Q-učenje, PPO i A2C testiraju u okruženju igrice
Pong. Najbolji algoritam koristi se u implementaciji sučelja za igranje igrice Pong.

Abstract (english)

Reinforcement learning grows more popular by inventing new methods and with
more regular applications. The field is not new and development started long before the
field became part of the schools curriculums. The Pong game is a popular environment
for testing and evaluating reinforcement learning algorithms. This paper gives an over-
view of the history of reinforcement learning and explains some of the fundemental
algorithms responsible for development of algorithms that are more used today. It also
gives a review of popular deep reinforcement learning algorithms. Finally it tests few
of them, Q-learning, PPO, and A2C on the Pong game. Best of those three algorithms
is used in an implementation of an application for playing the Pong game.

• Keywords: Duboko podržano učenje; igrica Pong; Q-učenje; PPO; RLHF
• Keywords (english): Deep reinforcement learning; Pong game; Q-learning; PPO; RLHF
• Language: croatian
• URN:NBN: urn:nbn:hr:168:283428
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2023-10-11 10:42:57