MARBLE: Multi-Armed Restless Bandits in Latent Markovian Environment
This addresses the challenge of adapting to dynamic environments in decision-making systems, such as recommender systems, but is incremental as it builds on existing RMAB frameworks.
The paper tackles the problem of decision-making in nonstationary environments by introducing MARBLE, a model that augments Restless Multi-Armed Bandits with a latent Markov state, and shows that their Q-learning method converges to an optimal policy, validated on a recommender system simulator.
Restless Multi-Armed Bandits (RMABs) are powerful models for decision-making under uncertainty, yet classical formulations typically assume fixed dynamics, an assumption often violated in nonstationary environments. We introduce MARBLE (Multi-Armed Restless Bandits in a Latent Markovian Environment), which augments RMABs with a latent Markov state that induces nonstationary behavior. In MARBLE, each arm evolves according to a latent environment state that switches over time, making policy learning substantially more challenging. We further introduce the Markov-Averaged Indexability (MAI) criterion as a relaxed indexability assumption and prove that, despite unobserved regime switches, under the MAI criterion, synchronous Q-learning with Whittle Indices (QWI) converges almost surely to the optimal Q-function and the corresponding Whittle indices. We validate MARBLE on a calibrated simulator-embedded (digital twin) recommender system, where QWI consistently adapts to a shifting latent state and converges to an optimal policy, empirically corroborating our theoretical findings.