基于仿真优化的急诊医护人员配置与调度问题的研究

Emergency department (ED) is characterized by treating patients with severe and complex diseases, rescuing the patients and managing heavy tasks in the hospital. ED should provide patients with rapid treatment, but the medical staff scarcity in ED causes serious imbalance between supply and demand, making the longer patients waiting time. Based on the stochastic nature and overcrowding of ED, as the mathematical methods fail to effectively describe the uncertainty caused by multiple unexpected factors in ED, this project studies the theories and methods of the ED medical staff planning and scheduling based on simulation optimization. Taking into account the complexity of the interaction between various medical procedures in ED, the fluctuation of the number of patients in different time slots, the disparity of the treatment process of patients with different categories and priorities, and the uncertainty of the doctor's diagnosis and treatment speed, we aim to propose the methods based on simulation optimization which could improve the resource utilization, reduce labor costs as well as the patients waiting time. This study extends the mathematical methods that only consider single-scene, single-skill medical staff planning and scheduling to the simulation optimization methods that take the multi-scene, multi-skill scenarios into account, not only enriching and developing the theories and methods of medical staff resources planning and scheduling, but also providing a scientific method for our hospital managers to comprehensively improve the efficiency of ED operations and cost savings.

急诊科是医院急危重患者多、病种复杂、抢救和管理任务重的科室。急诊本应为患者提供快速的救治，但急诊医护人员流失现象严重导致供需不平衡，使得患者候诊时间过长。本项目以急诊具有随机性且十分拥挤的现实问题为背景，针对数学解析方法难于解决多种不确定性相互影响的调度问题，研究基于仿真优化的急诊医护人员配置与调度的理论和方法。综合考虑急诊各医疗环节相互协作影响的复杂性、不同时段就诊患者数量的波动性、不同类型和等级患者就诊流程及优先级的非一致性、医生诊疗时间的不确定性等多场景因素，建立以提高资源利用率、降低人力成本和降低患者等待时间为目标的医护人员配置与调度问题的仿真优化方法。该项研究将原有仅考虑单场景、单技能医护人员配置与调度的解析方法拓展到考虑多场景、多技能的仿真优化方法，不仅丰富和发展了医疗人力资源配置和调度领域的理论与方法，也对我国医院管理者全面提高急诊运作效率、节约成本提供了科学方法。

本项目以急诊具有随机性且十分拥挤的现实问题为背景，针对数学解析方法难于解决多种不确定性相互影响的调度问题，研究基于仿真优化的急诊医护人员配置与调度的理论和方法。主要研究内容包括：（1）考虑时变患者需求的医护人员配置方法研究。即，根据不同时段不同类型患者的需求，动态配置医护人员，达到在有限资源的前提下使得患者等待时间最低的目的。本项目将该问题提炼为一个复杂的随机目标协同随机约束的离散仿真优化问题。（2）针对该问题的随机性与耗时性，本项目先后提出了混合保真度模型和基于高斯过程的具有自适应性的随机搜索算法，在大幅度缩减仿真时间的同时确保算法具有良好的鲁棒性。此外，（3）本项目还提出了一个在服务时间不确定下考虑医护之间默契程度的两阶段鲁棒优化模型。研究成果完善了现有的医疗资源配置理论，丰富了资源调度优化方法。成果包括发表SCI收录期刊学术论文11篇（第一作者或通讯作者论文7篇）。项目组成员中3位同学获得工学硕士学位，2名硕士研究生在读。此外，项目负责人还培养了三名优秀的本科毕业生，均在继续深造中。其中一名免试推荐到大连理工大学继续深造、一名获得Erasmus全额奖学金攻读硕士研究生。综上所述，课题组按照研究计划如期执行，达到了申报书所提出的目标和要求。