Mastering Complex Control in MOBA Games with Deep Reinforcement Learning

TL;DR

This paper introduces a deep reinforcement learning framework tailored for complex MOBA game control, achieving superhuman performance by handling large state-action spaces with novel algorithms and scalable system design.

Contribution

It presents a scalable RL system and novel algorithms like control dependency decoupling and dual-clip PPO for effective training in complex MOBA environments.

Findings

Tencent Solo defeats top professional players in 1v1 MOBA matches.

The proposed methods outperform existing RL approaches in complex game settings.

System design enables efficient large-scale exploration and training.

Abstract

We study the reinforcement learning problem of complex action control in the Multi-player Online Battle Arena (MOBA) 1v1 games. This problem involves far more complicated state and action spaces than those of traditional 1v1 games, such as Go and Atari series, which makes it very difficult to search any policies with human-level performance. In this paper, we present a deep reinforcement learning framework to tackle this problem from the perspectives of both system and algorithm. Our system is of low coupling and high scalability, which enables efficient explorations at large scale. Our algorithm includes several novel strategies, including control dependency decoupling, action mask, target attention, and dual-clip PPO, with which our proposed actor-critic network can be effectively trained in our system. Tested on the MOBA game Honor of Kings, our AI agent, called Tencent Solo, can defeat top professional human players in full 1v1 games.