Multi-agent coordination improves hospital patient flow and reduces delays
Physics-Informed Multi-Agent Coordination for Hospital Patient Flow Optimization
Artificial IntelligenceMultiagent Systems
Summary
Hospitals often face overcrowding and have many departments that need to work together to move patients efficiently. The authors combine mathematical models of queues with a type of artificial intelligence where many agents learn and negotiate from limited local information. Their method helps hospital departments coordinate patient routing and resource use dynamically, addressing real-world changes better than older approaches. Tests using real patient data show this teamwork-based approach lowers delays without risking patient safety.
What this means in practice
- •For hospital operations teams: Coordinate patient transfers and manage service capacity dynamically across departments to reduce wait times and balance workload.
- •For healthcare it system developers: Implement multi-agent AI systems that respect decentralized governance and local observations for real-time patient flow optimization.
Tested on one dataset.
Authors
Guoqing Zhang, Rafik Hadfi, Takayuki Ito
Abstract
Efficient patient flow coordination across autonomous hospital departments is critical for mitigating overcrowding and balancing resource utilization. While classical queueing theory, specifically open Baskett--Chandy--Muntz--Palacios (BCMP) networks, provides an interpretable mathematical topology for healthcare operations, analytical models rely on stationary assumptions and fixed routing matrices that degrade under state-dependent real-world dynamics. Conversely, centralized reinforcement learning approaches struggle to accommodate the decentralized structure of hospital governance, where individual clinical departments function with localized observations, heterogeneous resources, and divergent operational objectives. In this paper, we present a Multi-Agent Systems (MAS) framework titled \emph{Physics-Informed Multi-Agent Coordination}, which embeds empirically calibrated BCMP queueing topologies as physical priors within a decentralized multi-agent reinforcement learning architecture. Formulated as a Decentralized Partially Observable Markov Decision Process (Dec-POMDP) under coupled resource constraints, our method enables autonomous departmental agents to cooperatively negotiate patient routing and dynamic service scaling. To mitigate environmental non-stationarity without inducing excessive communication overhead, agents exchange localized action fingerprints along network edges and optimize a spatially decomposed reward structure. Empirical evaluations driven by real-world MIMIC-IV patient trajectories indicate that this cooperative multi-agent approach substantially reduces cumulative system delay compared to static Markovian approximations, heuristic dispatching, and independent multi-agent baselines, while maintaining clinical safety constraints.