GH Owl.Learning - Reinforcement Learning

,

Hi everyone,
I plan to use reinforcement learning within GH for an upcoming project and stumbled across Owl.learning by Mateusz Zwierzycki. Unfortunately, I couldn´t figure out how to set it up in the right way.

Anyone here who already used Owl Reinforcement Learning and could share a screenshot?

Thanks and all the best,
Emanuel

Welcome on board. You can @MateuszZwierzycki and he might come hooting :owl:

1 Like

QLearn.gh (29.5 KB)
Here is an example. Take a look at the reward calculation, you’ll see where the agent tries to get (it’s node 1)

2 Likes

Thanks a lot, Mateusz. It’s an amazing help to have your example!!
Regarding the reward calculation: if the “Next Step” equals “1” the reward output is True and the next state is randomly reset. I guess the reward has to not only reward the last action but, declining by step distance, the whole sequence of actions. What I couldn´t figure out is where/how the action sequence is saved until it gets rewarded, could you give me a hint?

thx for programming and sharing owl!