Minimax and Monte Carlo Tree Search Implementations for Two Player Game (CROSBI ID 710827)
Prilog sa skupa u zborniku | sažetak izlaganja sa skupa | međunarodna recenzija
Podaci o odgovornosti
Rončević, Toma ; Rodić, Marina; Despalatović, Ljiljana
engleski
Minimax and Monte Carlo Tree Search Implementations for Two Player Game
In this paper, we examine practical implementations of two algorithms: minimax and Monte Carlo Tree Search (MCTS). Both algorithms have been applied in the domain of two-player games with different extensions, modifications, and success. In this work, the two algorithms are implemented for a game that is an upgradeof the children's Tic-Tac-Toe game, now played on nine 3x3 boards with additional rules, called Ultimate Tic-Tac-Toe (UTTT). While the original game is trivial, this version is much harder to solve and represents a good use case for more advanced algorithms' implementation. The game was also used in an open competition among several hundred enthusiasts where our minimax approach arrived sixth in the finals. We give an overview of both algorithms, discuss their advantages and drawbacks, and describe their common modifications. We also discuss differences between algorithms, their results, and implementation details, both related and unrelated to our use case.
minimax ; mcts ; uttt ; implementation
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
Podaci o prilogu
521-521.
2021.
objavljeno
Podaci o matičnoj publikaciji
Proceedings of the 5th International Scientific and Professional Conference "The Challenges of Today"
Šibenik: Polytechnic of Šibenik
978-953-7566-46-3
Podaci o skupu
5. međunarodna znanstvena i stručna konferencija “IZAZOVI DANAŠNJICE” 5th International Scientific and Professional Conference “THE CHALLENGES OF TODAY”
ostalo
07.10.2021-08.10.2021
Šibenik, Hrvatska