Neighborhood Cognition Consistent Multi-Agent Reinforcement Learning
Abstract
Social psychology and real experiences show that cognitive consistency plays an important role to keep human society in order: if people have a more consistent cognition about their environments, they are more likely to achieve better cooperation. Meanwhile, only cognitive consistency within a neighborhood matters because humans only interact directly with their neighbors. Inspired by these observations, we take the first step to introduce \emph{neighborhood cognitive consistency} (NCC) into multi-agent reinforcement learning (MARL). Our NCC design is quite general and can be easily combined with existing MARL methods. As examples, we propose neighborhood cognition consistent deep Q-learning and Actor-Critic to facilitate large-scale multi-agent cooperations. Extensive experiments on several challenging tasks (i.e., packet routing, wifi configuration, and Google football player control) justify the superior performance of our methods compared with state-of-the-art MARL approaches.
Cite
@article{arxiv.1912.01160,
title = {Neighborhood Cognition Consistent Multi-Agent Reinforcement Learning},
author = {Hangyu Mao and Wulong Liu and Jianye Hao and Jun Luo and Dong Li and Zhengchao Zhang and Jun Wang and Zhen Xiao},
journal= {arXiv preprint arXiv:1912.01160},
year = {2020}
}
Comments
Accepted by AAAI2020 with oral presentation (https://aaai.org/Conferences/AAAI-20/wp-content/uploads/2020/01/AAAI-20-Accepted-Paper-List.pdf). Since AAAI2020 has started, I have the right to distribute this paper on arXiv