Penerapan Adaptive AI pada Game Turn Based RPG Dengan Menggunakan Metode Monte Carlo Tree Search

Yuka Bimatara Putra (Fakultas Ilmu Komputer, Universitas Brawijaya)
Eriq M. Adams Jonemaro (Fakultas Ilmu Komputer, Universitas Brawijaya)
Muhammad Aminul Akbar (Fakultas Ilmu Komputer, Universitas Brawijaya)

Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer
Vol 2 No 7 (2018): Juli 2018

Many methods can be used by game developers to be implemented into Artificial Intelligence (AI) in enemy agents or commonly called Non Playable Character (NPC) which aims to make players become more challenged in completing the game they play. One of the usual algorithms is Monte Carlo Tree Search (MCTS). The MCTS algorithm has been successfully constructed for turn-based gaming board games, the GO game. In the experiment, AI agents using MCTS scored higher than previous studies, so when tested, AIs using the MCTS algorithm can beat international GO champions. Given the success of the MCTS implementation, this study will discuss the application of MCTS to enemy agents in turn-based RPG games. The test is performed to validate the decision making of the selected agent and the effect of the MCTS process loop on the skill to be selected. For validation of decision making, testing will be done by looking at the effectiveness level and for testing the effect of MCTS process loop on effectiveness is done by simulation with several scenarios using different MCTS loops. The level of effectiveness is tested by looking at the win ratio of the agent team.