Institutional Repository of Vilnius University
Real-time game agent using monte carlo tree search /
Abstract
dc:descriptionThis work investigates the application of Monte Carlo tree search (MCTS) to a real-time, high FPS Bomberman-like game, focusing on challenges such as continuous movement and simultaneous actions. The research implements a single-player MCTS tree, where opponent behaviour is simulated using predefined strategies (static, walking, or bomb-placing agents). Key enhancements include a distance-based selection heuristic to prioritize navigation toward the opponent. The agent's performance is evaluated against rule-based opponents, investigating which opponent simulation strategies let the agent perform better, with results showing high win rates against more deterministic opponents and limitations against stochastic opponents. This work demonstrates how MCTS can be adapted to real-time games via a single-player tree with opponent simulation.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2025
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Budrys, Benas,
Subjects
dc:subject × 1Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- eng
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD239029982&prefLang=en_US