Hierarchical Reinforcement Learning for Multi-agent MOBA Game

  • 2019-05-31 03:00:05
  • Zhijian Zhang, Haozheng Li, Luo Zhang, Tianyin Zheng, Ting Zhang, Xiong Hao, Xiaoxin Chen, Min Chen, Fangxu Xiao, Wei Zhou
  • 0


Real Time Strategy (RTS) games require macro strategies as well as microstrategies to obtain satisfactory performance since it has large state space,action space, and hidden information. This paper presents a novel hierarchicalreinforcement learning model for mastering Multiplayer Online Battle Arena(MOBA) games, a sub-genre of RTS games. The novelty of this work are: (1)proposing a hierarchical framework, where agents execute macro strategies byimitation learning and carry out micromanipulations through reinforcementlearning, (2) developing a simple self-learning method to get better sampleefficiency for training, and (3) designing a dense reward function formulti-agent cooperation in the absence of game engine or ApplicationProgramming Interface (API). Finally, various experiments have been performedto validate the superior performance of the proposed method over otherstate-of-the-art reinforcement learning algorithms. Agent successfully learnsto combat and defeat bronze-level built-in AI with 100% win rate, andexperiments show that our method can create a competitive multi-agent for akind of mobile MOBA game {\it King of Glory} in 5v5 mode.


Quick Read (beta)

loading the full paper ...