.png){kind=icon}
返回
Structural relational inference actor-critic for multi-agent reinforcement learning
.png){kind=icon}
OA
AI
DOI：10.1016/j.neucom.2021.07.014.png)
.png){kind=icon}
.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Multi-agent reinforcement learning (MARL) is essential for a wide range of high-dimensional scenarios and complicated tasks with multiple agents. Many attempts have been made for agents with prior domain knowledge and predefined structure. However, the interaction relationship between agents in a multi-agent system (MAS) in general is usually unknown, and previous methods could not tackle dynamical activities in an ever-changing environment. Here we propose a multi-agent Actor-Critic algo-rithm called Structural Relational Inference Actor-Critic (SRI-AC), which is based on the framework of centralized training and decentralized execution. SRI-AC utilizes the latent codes in variational autoen-coder (VAE) to represent interactions between paired agents, and the reconstruction error is based on Graph Neural Network (GNN). With this framework, we test whether the reinforcement learning learners could form an interpretable structure while achieving better performance in both cooperative and com-petitive scenarios. The results indicate that SRI-AC could be applied to complex dynamic environments to find an interpretable structure while obtaining better performance compared to baseline algorithms. (c) 2021 Elsevier B.V. All rights reserved.
Keyword:
Multi-agent systems
Deep reinforcement learning
Variational autoencoder
Actor-critic
Graph neural network
AI总结
对已上传原文的论文进行重点信息的提取，主要内容包括：简要概述、研究摘要、背景介绍、关键亮点、图文解析、展望与总结。
期刊
IF:
6.5
论文数:
2.5W
被引数:
6.5W
机构
引用论文
.png){kind=icon}
分享收藏Study of bi-directional buck-boost converter topologies for application in electrical vehicle motor drives应用于电动汽车电机驱动的双向buck-boost变换器拓扑研究
PREAI
分享收藏分享收藏
查看更多内容