Abstract—Ms. Pac-Man has been a very popular arcade game since its release in 1982. The original game is based on a single maze structure which can make game play lose its attractiveness eventually. This paper aims to generate sets of various mazes through an evolutionary approach. A genetic algorithm was designed to create optimal mazes by specifying a fitness function to create different mazes which can allow the game to be finished by the player. Evolutionary approach was chosen due to its unique yet acceptable results. Results show that different maze structures are possible to obtain in addition to the classical design.
Index Terms—Genetic algorithm, evolutionary maze generation, Ms. Pac-Man, artificial intelligence.
The authors are with SAAT Lab in Computer Engineering Department, Ankara University, Turkey (e-mail: a.buraksafak@gmail.com, ebostanci@ankara.edu.tr, alisoylucicek@gmail.com).
Cite: Aykut Burak Safak, Erkan Bostanci and Ali Emre Soylucicek, "Automated Maze Generation for Ms. Pac-Man Using Genetic Algorithms," International Journal of Machine Learning and Computing vol. 6, no. 4, pp. 226-230, 2016.