The creation of an AI in a procedural environment aims to develop both an environment and artificial intelligence (AI) for non-playable characters (NPCs) in a zombie genre video game. The project aims to explore certain gaps in the artificial intelligences of this genre, seeking new ways to provide an immersive experience where characters act autonomously and interact with each other in order to survive.

For the development of this project, mechanics and algorithms that are efficient for optimal and effective production are deeply analyzed, and various video games related to the established goals are also compared. In addition to AI, the creation of a procedural environment where NPCs can operate with complete freedom is also proposed as an objective. All the necessary resources to achieve the project's goals were also modeled.

The initial steps taken inside the project were related to the procedural generation of the world.

The next step was to begin developing artificial intelligence and its fundamental systems. Each of the NPCs will be randomly generated at the beginning of each game session. The project's intention with this is to promote replayability and offer a different gaming experience every time a new game is created. Specifically regarding the NPCs, they must have various essential aspects that set them apart, including personal and physical characteristics.

The attribute system will define the characters' personalities, influencing how they interact with each other. In a post-apocalyptic world, survivors' skills will be based on their former professions. A navigation system will be implemented, affecting character movement and influenced by attributes such as fatigue. Additionally, a relationship system will be based on the characters' morality, which will evolve during the game. Relationships between characters will be formed based on various mentioned systems and attributes. Lastly, task systems will be established, dependent on the needs of survivors and their groups.

These are some GIFs that show the final outcome.