About the project
This was a university project that I had the honor of making. The project was you have to make AI in a nonconventional way such as behavior trees or state machines so not just a bunch of if statements.
​
In this project I will be creating a AI that will be in control of an evolution system with prey and predator aspects such as when they reproduce the offspring will have genes that will get passed on such as being faster or longer vision etc. I will be picking the pathfinding contract, with some aspects of adaptive difficulty system this is because with the idea I am going for the AI will need to move around the map with some advanced movement such as steering around on its own and making decisions on where to go by objects being in the way or going for objects they need like food. Another way its needs to have advanced movement is so that the AI can get away from predators so which means out running them or hiding. And the adaptive system comes into it by taking factors in to account like if how hungry the pray is and how many predators there are and then the prey will decide of its worth getting food at that time due to the number of predators. This all goes for predator as well but will have different parameters.
Lotka-Volterra model where when one population goes down the other one goes up then vice versa and that was my goal for the project.
​
So the way I created the AI was I used fuzzy logic where it's not just yes or no there would be more give and take on how the AI will move to what extent the number gets back from Raycasts from the AI
​
​
This is how the AI sees it uses raycasts so that it can detect whats around it such as food predators or prey and mates so that it can reproduces
The script above is where I can control the gene mutations and also what genes get inherited from which parent depending on whos the stronger one and which ones are best for the offspring. The mutation would only be on a rare.
This is the poster I created on the project on what research I did and also on the data I collected so over all where I wanted to get my data to represent the Lotka-Volterra model I feel like I did it.