Implementing and testing an AI using the Monte Carlo Tree Search method for imperfect information trick-tacking games, like Witches and Wizard[go to overview]
Artificial Intelligence agents are as old as video games themselves, with the first appearing around the 1940s. These AIs try to simulate responsive, adaptive and intelligent behaviour mostly in NPCs (Non-Player-Characters), to let them appear human like, with the aim of improving the game playexperience of the human players. As games today become bigger, more realistic and more complicated, more and more reliable and efficient algorithms and methods are needed to ensure the ability of humanlike realtime decision making. The goal of this thesis is to further evaluate the Monte Carlo Tree Search method for imperfect information games, like "Witches" and "Wizard".
22.10.20 - 10:15
via Big Blue Button