Skip to content
MAGICS Lab

ASME Journal of Computing and Information Science in Engineering · 2025

Adaptive Network Intervention for Complex Systems: A Hierarchical Graph Reinforcement Learning Approach

Hierarchical RL that steers an evolving multi-agent network toward cooperation.

Qiliang Chen, Babak Heydari

Hierarchical Graph Reinforcement Learning system diagram: HGRL framework observes and intervenes in an evolving network of autonomous agents that update their types by imitation.
Hierarchical Graph Reinforcement Learning system diagram: HGRL framework observes and intervenes in an evolving network of autonomous agents that update their types by imitation.

How can a system manager with limited authority steer a learning multi-agent system toward cooperation when the network keeps evolving? This paper introduces a Hierarchical Graph Reinforcement Learning (HGRL) framework that intervenes in network structure to promote pro-social behavior.

The hierarchy collapses what would otherwise be an unmanageable state and action space. HGRL beats established baselines, and the resulting networks tend toward robust core-periphery structures dominated by cooperators — especially when peer-to-peer social learning is weak.