Artificial Intelligence Based Game Levelling

Authors : Meric CETİN, Yunus SARICA

Pages : 147-153

Doi:10.17694/bajece.642973

View : 14 | Download : 8

Publication Date : 2020-04-30

Article Type : Research Paper

Abstract :The game development process is becoming a more detailed structure every day. The applications of artificial intelligence insert ignore into journalissuearticles values(AI);, which is a comprehensive information technology, have been closely related to game technologies. In this study, the levelling process of a 2-dimensional insert ignore into journalissuearticles values(2D); platform game was investigated. The game developed and called “ Renga ” has a basic gameplay. Game data has been processed through an artificial neural network insert ignore into journalissuearticles values(ANN);, k -nearest neighbour, decision and random tree algorithms and deep learning model that is trained with gameplay and user information. The classification process with the output data provides results for the next game level. In this way, the most effective playability impression that the developers offer to the game users has been created according to game. Furthermore, the variety of difficulty calculated with dynamic data by the user is provided by Renga , in which new sections/levels are created with user-specific assets. Thus, the most efficient gaming experience has been transferred to the users.
Keywords : Artificial Intelligence, Difficulty Adjustment, Content Generation, k Nearest Neighbor, Random Forest, Artificial Neural Networks