Procedural Content Generation of Custom Tower Defense Game Using Genetic Algorithms

In the present day, it is difficult to imagine the development of computer games without the use of artificial intelligence. We see it utilized for gameplay, players modeling, playtesting, or content generation. In this paper, we focused on the content generation of a custom Tower Defense game named Save the Sheep. The Tower defense game is a strategic game, which was, in our case, proposed in a non-violent way. We generated key building blocks of the game with a genetic algorithm, i.e., a game map, unit placement, and a waves system. The final Tower Defense inspired game was implemented in the Unity game engine. The results show that by applying genetic algorithms, it is possible not only to generate content that makes the game more complex, but also more challenging and interesting for players.