Towards Heterogeneous Multi-Agent Reinforcement Learning with Graph Neural Networks
This work addresses the challenge of handling heterogeneous entities in multi-agent systems, which is an incremental improvement in a domain-specific area.
The paper tackles the problem of learning policies for multiple agent classes in heterogeneous multi-agent reinforcement learning by proposing a neural network architecture that uses directed labeled graphs and relational graph convolution layers to specialize communication channels between entity types, resulting in higher performance in such environments.
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.