Article: Dynamic scheduling of hospital social security settlement based on multi-agent reinforcement learning Journal: International Journal of Information and Communication Technology (IJICT) 2025 Vol.26 No.43 pp.95 - 111 Abstract: Hospital social security settlement is a core link connecting medical services, medical insurance systems and patient interests, with its operational efficiency directly affecting medical service quality and social security system sustainability. Expanded medical insurance coverage, surging daily settlements and frequent policy adjustments make traditional static scheduling unable to adapt to system dynamics, causing long patient waits, low terminal utilisation and high verification failures. This study proposes a dynamic scheduling framework for hospital social security settlement based on multi-agent reinforcement learning, with four intelligent agents for distributed decision-making, plus a multi-objective reward function and constrained action mechanism. Experiments with real data from a tertiary Grade A hospital show the framework cuts average settlement delay by 38.2% and 21.5%, raises terminal utilisation by 27.6% and maintains over 99.5% compliance. It offers an intelligent solution to boost settlement efficiency and supports medical insurance service digital transformation. Inderscience Publishers - linking academia, business and industry through research

Title: Dynamic scheduling of hospital social security settlement based on multi-agent reinforcement learning

Authors: Yinping Tian

Addresses: Heping Hospital, Changzhi, 046000, China; Affiliated to: Changzhi Medical College, China

Abstract: Hospital social security settlement is a core link connecting medical services, medical insurance systems and patient interests, with its operational efficiency directly affecting medical service quality and social security system sustainability. Expanded medical insurance coverage, surging daily settlements and frequent policy adjustments make traditional static scheduling unable to adapt to system dynamics, causing long patient waits, low terminal utilisation and high verification failures. This study proposes a dynamic scheduling framework for hospital social security settlement based on multi-agent reinforcement learning, with four intelligent agents for distributed decision-making, plus a multi-objective reward function and constrained action mechanism. Experiments with real data from a tertiary Grade A hospital show the framework cuts average settlement delay by 38.2% and 21.5%, raises terminal utilisation by 27.6% and maintains over 99.5% compliance. It offers an intelligent solution to boost settlement efficiency and supports medical insurance service digital transformation.

Keywords: hospital social security settlement; dynamic scheduling; multi-agent reinforcement learning; MARL; intelligent agent; settlement efficiency.

DOI: 10.1504/IJICT.2025.150401

International Journal of Information and Communication Technology, 2025 Vol.26 No.43, pp.95 - 111

Received: 28 Sep 2025
Accepted: 30 Oct 2025
Published online: 12 Dec 2025 *

Title: Dynamic scheduling of hospital social security settlement based on multi-agent reinforcement learning

Keep up-to-date