Small sample of Q-Learning in C.
The (A)ctor must maximize the score, which means it should avoid the (T)raps, reach the (O)bjective and, if possible, pick (B)onus objects.
The Actor does not know anything about the map and cannot see anything, it doesn't use graphs. It just tries a lot of times, starting with zero knowledge, and improves his behavior through try and error.
Execute make
in current directory.
`./reinforcement-learning
C is probably not the best language to do reinforcement learning. This is only a hobbie project.