Developing a Successful Bomberman Agent
This work addresses the challenge of creating competitive AI for a specific game domain, representing an incremental improvement in game-playing agents.
The paper tackled the problem of developing an AI agent for a Bomberman variant on the CodinGame platform, achieving top performance by ranking first among 2,300 submitted agents using a Beam Search approach with enhancements like bit-based state representation and simulation-based pruning.
In this paper, we study AI approaches to successfully play a 2-4 players, full information, Bomberman variant published on the CodinGame platform. We compare the behavior of three search algorithms: Monte Carlo Tree Search, Rolling Horizon Evolution, and Beam Search. We present various enhancements leading to improve the agents' strength that concern search, opponent prediction, game state evaluation, and game engine encoding. Our top agent variant is based on a Beam Search with low-level bit-based state representation and evaluation function heavy relying on pruning unpromising states based on simulation-based estimation of survival. It reached the top one position among the 2,300 AI agents submitted on the CodinGame arena.