This repository contains the work of the group 10 for the final project of Learning Dynamics. It is a tentative of replication of the paper Silver, D., & Veness, J. (2010). Monte-Carlo planning in large POMDPs. Neural Information Processing Systems.
To make our project work, one must import the library https://github.com/d3sm0/gym_pomdp where we applied some modifications on the environment rock and battleship. Those changes can be found on the files Rock2.0.py and Battleship2.0.py. To experiments with our POMCP implementation, the user must use the files RockGame.py or BattleshipGame.py depending on the game to use as environment.