izpis_h1_title_alt

Učenje igranja realno-časovne strateške igre z uporabo globokega spodbujevalnega učenja
HABJAN, JERNEJ (Author), Guid, Matej (Mentor) More about this mentor... This link opens in a new window, Šter, Branko (Co-mentor)

.pdfPDF - Presentation file, Download (9,14 MB)

Abstract
Z algoritmom AlphaZero smo implementirali učenje in priporočanje akcij v realno-časovni strateški igri. Pregledali smo krajšo zgodovino globokega spodbujevalnega učenja na igrah in povzeli, zakaj je pristop samostojnega učenja najprimernejši. Za strateško igro smo določili stanja igre in stanje s kodirnikom preoblikovali v format, primeren za učenje nevronske mreže. Določili smo ustavitveni pogoj z iztekom števila preostalih potez za posameznega igralca. Utemeljili smo različne konfiguracije učnih parametrov in izpostavili tisto, ki računalniškega agenta na naši igri najuspešnejše uči. Rezultate smo prikazali s Python knjižnico Pygame in v celostnem pogonu Unreal Engine 4. V obeh vizualizacijah lahko igramo proti naučenemu modelu ali opazujemo, kako se dva računalniška nasprotnika bojujeta med sabo.

Language:Slovenian
Keywords:AlphaZero, realno-časovna strateška igra, Unreal Engine
Work type:Bachelor thesis/paper (mb11)
Organization:FRI - Faculty of computer and information science
Year:2019
Views:301
Downloads:128
Metadata:XML RDF-CHPDL DC-XML DC-RDF
 
Average score:(0 votes)
Your score:Voting is allowed only to logged in users.
:
Share:AddThis
AddThis uses cookies that require your consent. Edit consent...

Secondary language

Language:English
Title:Learning to play a real-time strategy game with deep reinforcement learning
Abstract:
With algorithm AlphaZero we have implemented the learning and recommendation of actions in a real-time strategy game. We examined a short history of deep reinforcement learning in games and summarized why the self-learning approach is best suited. For a strategic game, we determined the state of the game and transformed it with the encoder into a format suitable for learning a neural network. We determined a stopping condition with the expiry of the number of remaining moves for each player. We substantiated different configurations of learning parameters and exposed the most successful configuration for learning our game. The results were displayed with the Python Pygame module and the game engine Unreal Engine 4. In both visualizations we can play against the learned model, or we can observe two computer opponents fighting against each other.

Keywords:AlphaZero, real-time strategy game, Unreal Engine

Similar documents

Similar works from RUL:
Similar works from other Slovenian collections:

Comments

Leave comment

You have to log in to leave a comment.

Comments (0)
0 - 0 / 0
 
There are no comments!

Back