There are several approaches of applying artificial neural networks (ANN) as game AI. Most of these experiments deal with evolving networks to control player characters (bot players) or a specific non-playable character (NPC), furthermore, we have seen attempts to create adaptive AI for NPC's that adjust actions with player actions and environmental conditions. The main goal here is to create a simulation with different species of NPC's that adapt to changes, players and each other.
The task for the semester is to plan and implement a game that roughly simulates an ecosphere, where there are different species like plants, herbivore/carnivore animals and a human as a playable character. The animals have to eat and rest to recover and plants have to grow and reproduce. The student also has to create some basic AI for the different species that will be later replaced with some ANN.