智能系统学报2025，Vol.20Issue(4)：800-812,13.DOI:10.11992/tis.202404032
基于机器学习的线上线下联合服务模式下医生排班算法
Algorithms for physician scheduling under the online and offline combined service mode based on machine learning
摘要
Abstract
The online and offline combined medical service mode has become a new medical service mode generally ad-opted by large hospitals in China.Under this mode,large hospitals need to allocate physicians to online and offline ser-vices,and arrange online and offline scheduling plans for physicians while considering the switching of physicians between the two services.To address this problem,a Markov decision process model for physician scheduling with ser-vice level constraints was developed and an approximate dynamic programming algorithm was designed to solve the Markov decision process with high efficiency.Furthermore,considering multi-dimensional uncertainties such as highly time-varying patient arrival and service hours,a data-driven recurrent neural network was constructed based on the real-life data of the cooperative hospital as a performance evaluation method for the online and offline queueing systems.Numerical experiments show that the proposed methods can reduce the total working hours of physicians,effectively control the waiting time of patients,and ensure the high-efficiency operation of the system.关键词
线上医疗/医生排班/时变排队系统/数据驱动/深度学习/马尔可夫决策过程/近似动态规划/启发式算法Key words
telemedicine/physician scheduling/time-varying queueing system/data-driven/deep learning/Markov de-cision process/approximate dynamic programming/heuristic分类
信息技术与安全科学引用本文复制引用
张越,王子翔,周博,刘冉,杨之涛..基于机器学习的线上线下联合服务模式下医生排班算法[J].智能系统学报,2025,20(4):800-812,13.基金项目
国家自然科学基金项目(72371161) （72371161）
上海申康医院发展中心管理研究项目(2024SKMR-19) （2024SKMR-19）
上海交通大学中国医院发展研究院研究项目(CHDI-2024-A-04). （CHDI-2024-A-04）
