Causality Meets Locality: Provably Generalizable and Scalable Policy Learning for Networked Systems
This work addresses scalability and generalization issues for reinforcement learning agents in large-scale networked systems like traffic and power grids, offering a novel approach with theoretical backing.
The paper tackles the challenge of scaling reinforcement learning to large networked systems while handling environment shifts, by proposing GSAC, a framework that combines causal representation learning with meta actor-critic learning, achieving rapid adaptation and outperforming baselines with finite-sample guarantees.
Large-scale networked systems, such as traffic, power, and wireless grids, challenge reinforcement-learning agents with both scale and environment shifts. To address these challenges, we propose GSAC (Generalizable and Scalable Actor-Critic), a framework that couples causal representation learning with meta actor-critic learning to achieve both scalability and domain generalization. Each agent first learns a sparse local causal mask that provably identifies the minimal neighborhood variables influencing its dynamics, yielding exponentially tight approximately compact representations (ACRs) of state and domain factors. These ACRs bound the error of truncating value functions to $κ$-hop neighborhoods, enabling efficient learning on graphs. A meta actor-critic then trains a shared policy across multiple source domains while conditioning on the compact domain factors; at test time, a few trajectories suffice to estimate the new domain factor and deploy the adapted policy. We establish finite-sample guarantees on causal recovery, actor-critic convergence, and adaptation gap, and show that GSAC adapts rapidly and significantly outperforms learning-from-scratch and conventional adaptation baselines.