The project scope is to implement a Bomberman agent with help of reinforcement learning in order to compete in a competition amongst students. The Bomberman environment framework is based on https://github.com/ukoethe/bomberman_rl.
In the directory agent_code
are several rule-based and random agents from https://github.com/ukoethe/bomberman_rl as well as our own reinforcement agents.
Over all three agent approaches have been implemented: noodle_gen_1
, noodle_gen_2
, noodle_gen_3