Shameless self plug: if you're interested in Markov Decision Processes and related algorithms
, I am the author and maintainer of a C++ and Python library that implements lots of planning and reinforcement learning methods for them.
I hope to officially publish it later this year, but has already been used by a lot of people, in academia and out.
(Author here.) That looks like a great library --- I will have to try it out!
In case it helps anyone, I also wrote a much smaller MDP library in ruby to support this 2048 project (and several others): https://github.com/jdleesmiller/finite_mdp . It basically just does value iteration and policy iteration.
Thank you too for putting the code for your article out there. There is always a shortage of public code to study this field, and every bit helps.
I've actually wanted to write a similar article on 2048 for a while, but unfortunately I've never found the time. But it seems now I won't have to, your article is great and much better than anything I could have ever done!
I hope to officially publish it later this year, but has already been used by a lot of people, in academia and out.
Feel free to play with it if you like!
https://github.com/Svalorzen/AI-Toolbox