Towards Heterogeneous Multi-Agent Reinforcement Learning with Graph Neural Networks

This work proposes a neural network architecture that learns policies for multiple agent classes in a heterogeneous multi-agent reinforcement setting. The proposed network uses directed labeled graph representations for states, encodes feature vectors of different sizes for different entity classes, uses relational graph convolution layers to model different communication channels between entity types and learns distinct policies for different agent classes, sharing parameters wherever possible. Results have shown that specializing the communication channels between entity classes is a promising step to achieve higher performance in environments composed of heterogeneous entities.


page 6

page 9


Heterogeneous Multi-Robot Reinforcement Learning

Cooperative multi-robot tasks can benefit from heterogeneity in the robo...

Learning Transferable Cooperative Behavior in Multi-Agent Teams

While multi-agent interactions can be naturally modeled as a graph, the ...

Multi-agent Reinforcement Learning with Graph Q-Networks for Antenna Tuning

Future generations of mobile networks are expected to contain more and m...

ModGNN: Expert Policy Approximation in Multi-Agent Systems with a Modular Graph Neural Network Architecture

Recent work in the multi-agent domain has shown the promise of Graph Neu...

Reward-Sharing Relational Networks in Multi-Agent Reinforcement Learning as a Framework for Emergent Behavior

In this work, we integrate `social' interactions into the MARL setup thr...

Applications of Multi-Agent Reinforcement Learning in Future Internet: A Comprehensive Survey

Future Internet involves several emerging technologies such as 5G and be...

Please sign up or login with your details

Forgot password? Click here to reset