/DeepLearning-Bomberman

Reinforcement Learning (DQN, PPO, MCTS+DQN) for the classical game Bomberman.

Primary LanguageJupyter Notebook

FML Project

This repository contains the code for an agent playing the classical game Bomberman. Three main approaches are implemented:

  • Deep Q Learning (DQN)
  • Proximal Policy Optimization (PPO)
  • Monte-Carlo Tree Search (MCTS) in conjunction with DQN