Reinforcement learning s aplikací ve hrách

Abstract

This thesis deals with the study of machine learning, specifically reinforcement learning algorithms and their application in the game. The game called checkers was created with chosen rules in the Unity game engine environment. Subsequently, two reinforcement learning algorithms were implemented in the same environment and then applied to a checkers game. The program was launched for artificial intelligence to learn how to play the game. Upon completion of the learning process, tests were designed to find out, how intelligent are learned artificial intelligences. These tests include both: artificial intelligence against artificial intelligence games and human against artificial intelligence games. At the end, the tests results were evaluated.