有限分布信息环境下手术调度问题的鲁棒优化方法研究

This project is aimed to research the operating room resources allocation and surgery scheduling problems with limited distribution information on service time and emergency arrival that addressing the practical needs of hospital surgery departments. Distributionally robust optimization is applied to deal with surgery scheduling problems considering different environments such as multiple-priority patients, different category patients and surgery cancellation. The Research contents include the surgery scheduling problem with stochastic service time and multiple-priority patients, the surgery scheduling problem with surgery cancellation and the surgery scheduling problem with stochastic emergency arrival time. The original surgery scheduling methods rely on historical data, while this research offers an attractive alternative for addressing uncertainty in optimization modeling rather than specifying exact probability distributions. And also the model is extended to consider stochastic emergency arrival. All the above considerations facilitate smoothing patients flow and resources, and improve the robustness of the method. The results could enrich and develop the surgery scheduling theory and methods, and provide technical supports for building the scientific operating room management system with high efficiency.

本项目以医院手术室运作管理环境为背景，针对手术服务时间不确定及调度方案不能过分依赖随机变量分布的现实需求，研究考虑手术服务时间和急诊患者到达为有限分布信息环境下的手术室资源分配策略及手术调度问题，建立针对不同场景（多类型、多优先级患者和手术停台）的手术室调度问题的分布鲁棒优化方法，包括:考虑随机服务时间的多优先级手术调度方法；考虑手术停台风险的手术调度方法；考虑急诊患者动态到达的手术调度方法。该研究将原有的过分依赖分布规律的手术室调度策略拓展到基于有限分布信息的手术室鲁棒调度模型，同时拓展到考虑急诊患者动态到达的手术调度模型，既增加了方法的抗扰动性，也有利于平衡资源和患者流的分配。研究成果不仅丰富医院手术排程的理论与方法，也对我国医院管理者建立科学、高效的手术室管理系统提供方法与技术支持。