基于MDP和动态规划的医疗检查预约调度优化方法研究

医疗检查对医生诊断病人病情具有重要作用。针对医疗检查资源的预约调度问题,考虑两台设备、三类病人且各类病人所需检查时间不同的情况。以医院在检查设备方面收益最大化为目标,建立有限时域马尔可夫决策(Markov decision process,MDP)模型,并结合动态规划理论,得出系统最优的预约排程策略。通过matlab仿真模拟医院的检查预约情况,并结合调研数据,实例验证了该预约策略相对于传统预约策略的优越性。最后,对设备的最大可用时间和住院病人的预约请求到达率模型进行敏感性分析,研究了预约策略的适用性。

Appointment Scheduling of Medical Examination Based on MDP and Dynamic programming

Medical examination plays an important role in the diagnosis of disease. In view of the appointment scheduling of medical examination equipment, this paper considers two sets of medical equipment and three types of patients (outpatients, inpatients, emergency patients) with different check time. With the goal of maximizing the revenue in medical examination system, a finite horizon Markov decision process (MDP) model is proposed. Then, combined with dynamic programming model and iteration, the optimal appointment scheduling strategy can be obtained. Using matlab to do the programming, this paper simulates the arrival of different patients. Based on the setting of related parameters, the numerical examples show that the scheduling strategy in this paper is superior to the traditional appointment strategy. Finally, sensitivity analysis is carried out to explore the applicability of this appointment scheduling strategy. The conclusion is that Markov decision process (MDP) is suitable for the appointment scheduling of medical examination equipment. When the system capacity is scarce, or the arrival rate of inpatients increases, this decision model will be superior to traditional strategy.