A Scalable Approach to Solving Simulation-Based Network Security Games
This provides a practical solution for efficient hierarchical policy learning in large-scale networked decision problems, such as cyber-network security, though it appears incremental by augmenting existing paradigms.
The paper tackles the challenge of scaling multi-agent reinforcement learning for network security games by introducing MetaDOAR, a meta-controller that uses learned filtering and caching to achieve higher player payoffs than state-of-the-art baselines on large network topologies without significant memory or training time issues.
We introduce MetaDOAR, a lightweight meta-controller that augments the Double Oracle / PSRO paradigm with a learned, partition-aware filtering layer and Q-value caching to enable scalable multi-agent reinforcement learning on very large cyber-network environments. MetaDOAR learns a compact state projection from per node structural embeddings to rapidly score and select a small subset of devices (a top-k partition) on which a conventional low-level actor performs focused beam search utilizing a critic agent. Selected candidate actions are evaluated with batched critic forwards and stored in an LRU cache keyed by a quantized state projection and local action identifiers, dramatically reducing redundant critic computation while preserving decision quality via conservative k-hop cache invalidation. Empirically, MetaDOAR attains higher player payoffs than SOTA baselines on large network topologies, without significant scaling issues in terms of memory usage or training time. This contribution provide a practical, theoretically motivated path to efficient hierarchical policy learning for large-scale networked decision problems.