Institutional Repository of Vilnius University
Catan Žaidimo Agento Kūrimas Naudojant Monte Karlo Medžių Paiešką /
Abstract
dc:descriptionThis study explores various methods for developing game agents, specifically in the context of the board game Settlers of Catan. Catan presents unique challenges for an artificial intelligence agent, as it is designed for more than two players, is non-deterministic, involves imperfect information, and features a vast number of possible game states. For such a game, Monte Carlo Tree Search (MCTS) has been proven to be the most suitable and thus became the primary method investigated. MCTS is a statistical algorithm widely used in AI decision-making tasks, including both deterministic games (like Chess and Go) and non-deterministic games (such as Catan, Poker, and various video games). Throughout this study, a broad range of MCTS applications and related methodologies were analyzed, with a particular focus on adapting MCTS for the development of a Catan game agent. For this research, Natak API - a C\# implementation of the Catan game, built upon the HTTP REST protocol - was utilized and modified to fit the needs of MCTS agent development. Interacting with this API provided the necessary framework for developing and testing the game agents. Within this program, three distinct agents were developed: a random agent capable of quickly playing games by performing arbitrary actions, a deterministic agent that plays according to specific predefined rules, and the MCTS agent, which applies the Monte Carlo Tree Search algorithm to its decision making.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2025
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Krikščiūnas, Simas,
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD239031292&prefLang=en_US