Here is an idea I have for a game that you may use. It uses artificial intelligence and neural whatever to evolve monsters that hunt after the player. (when it says the monsters that have the most score get the most points, i meant to say that the monsters with the highest points will breed)

_____



A.I. Expirements

Game




_____

A player is set in a standard survival game with materials to build a shelter and plenty of crafting




There are monsters that gain points for going towards the player. Here is the scoring:




They gain a little bit of points for being close to the player

They gain a tiny bit of points for attacking

They gain many points for damaging the player

They gain a lot of points for killing the player

They lose many points if the player hurts them

At the end of every day, the monster will attempt to go back to their base to breed. The highest scoring monsters gain the most points.




At the same time, the monster must keep itself alive by drinking water and eating food




_____

To make a variety of monster behavior, add in score variables like:




> lose less ( or gain more ) points while near another monster of the same species ( this will most likely form packs of monsters that attack the player at once )

> gain some points for attacking other predators ( to make rivalries )

> lose points for being near player or being attacked by player ( creates passive monsters that will not attack )

> lose points for leaving a specific area ( creates biomes where these monsters will stay in and gravitate towards )




_____

Monsters can perform many actions:




> grabbing objects (the larger the monster, the larger the objects they can grab)

> jump

> walk

> attack

> making sounds ( maybe will also give points )

> eat

> drink

> sleep




_____

Monsters will most likely be randomized