Hello Comunity,
i'm working on an artificial intelligence project just for fun and like to offer my code for experimenting. it is very simple and extentable.
What it does:This approach uses genetic algorithm for learning to find a path to the next food spot. In the define-section are the values, how many agents (animats, bots) to use, how many genes they can store to evaluate, the mutation rate and the actions they can take.
Don't expect a graphical high-end scenario. This is just the beginning. And my coding coudl be improved too, but this is my first attempt and i have to learn alot with pointer, structs, functions and their efficiency. for hints i'm thankfull.
What can you expect in the future:I have a goal in my mind, where the learning algorithm will be extended with reinforcment learning, fuzzy logic, finite state machine and rule-based decision making. But everything in a dynamic way. For example the GA is static in his form. The genomes shoudl be dynamic, that if a goal is reached, the agents can store the sequence they evolved at harddisk and later recall when needed. It is like episodic knowledge.
Then later it should included a short and longterm working memory, procedural, declarative, episodic und possibly iconic forms of knowledge to build a system on open-ended problemspaces. Interacting with the outside world and learn about all tasks and its performance on those tasks.
Yes, i know this is a long way to take, but i do my best to achieve it.
so have fun with the minimal code.
Download:
http://ul.to/tif4tyxsPriNova
