【作者】 王恪铭；

【导师】 马祖军；

【作者基本信息】 西南交通大学 ， 交通运输安全工程， 2013， 博士

【摘要】 各类具有严重影响与后果的非常规突发事件在全球时有发生,其中地震、特大恐怖袭击等通常会造成大量的人员伤亡。血液是一类特殊的救援物资,在医疗应急保障中发挥着至关重要的作用。为应对因突发事件造成大量伤员而引起的用血量增长,当灾区血站库存与采集量难以满足临床需求时,需要从非灾区血站紧急调剂。血液保障直接关系到受伤群众的生命安全,可见对非常规事件发生后血液应急调剂问题进行研究,具有十分重要的社会意义。在血液应急调剂的过程中,需要合理地进行出救血站的选择,并确定出各个出救血站分配至各灾区血站的各血制品的数量；优化多种运输方式联合的路线安排；当灾区血站库存出现过量增加的情形时,必须及时转运积压血液制品以减小潜在的过期报废；血液在应急分配中也不可避免地遇到短缺,需要结合各血型相互之间的替代关系,以优化提高分配方案的满意程度。本文主要研究内容如下：(1)通过分析血液特性与应急血液保障的阶段性特点,并考虑运输方式对血液质量的影响,建立了一种多阶段优化模型解决多灾点、多品种应急出救血站选择-分配问题,设计了一种基于关键受灾点优先的向量编码遗传算法进行求解,并对算法的时间复杂性进行了分析。最后,以“5.12”汶川大地震应急血液保障为背景构建算例进行验证,结果表明该遗传算法的计算效率较高,所提出的多阶段决策方法在血液新鲜度、血液积压量等方面具有显著优势。(2)基于血液的生理特性及应急保障特性,以应急血液最晚运达时间最小、接收时的最低新鲜度最大、运输总费用最小为目标,基于两阶段决策方法建立了考虑多品种、多式联运的应急出救血站选择-运输路线安排问题优化模型,并设计了一种包含局部邻域优化的遗传-禁忌混合算法进行求解。最后,以“5.12”汶川大地震应急血液保障为背景构建不同的血液需求情景,验证了模型与算法的有效性,并分析了目标函数权重系数与需求预测误差对调剂方案的影响。结果表明,多式联运可以提高血液调剂的及时性与效率；通过与单阶段决策方式在最晚抵达时间、新鲜度均值、运输总费用、血液积压量等指标上的对比,研究证实了两阶段决策方法的稳健性。(3)通过分析血液生理特性,确定出各血型相互之间的替代关系与优先级别,基于双层规划方法建立了考虑血型替代策略的多灾点、多品种应急血液分配问题优化模型,上层模型以最小化血液短缺量为目标,下层模型以最小化血型替代权重为目标。为便于求解,在保证最优解一致性的前提下,该问题被转化为单目标混合整数规划模型。推理了血型替代性质,设计了一种多阶段贪婪启发式算法进行求解,并对算法的时间复杂性进行了分析。最后,以汶川大地震应急血液保障为背景构建算例,验证了该模型的可行性,并分析了替代率对血液分配方案的影响,研究证实了采用血型替代策略可提高血液需求的满足率。通过不同规模的算例测试与CPLEX软件的求解结果对比,表明所提算法具有较高的适应性与计算效率。(4)在对血液库存状态与过期特性进行分析的基础上,基于库龄推导出血液转运、接收的相关性质,建立最小化运输费用的转运模型；最后通过算例仿真证实了文中推理与模型的有效性,研究验证了不同转运制品选择策略的差异以及保障概率、需求饱和度对使用新鲜度的调节作用,并分析了决策时刻、制品保质期对决策效果的影响,为应急保障中积压血液的转运决策问题提供了理论依据与实例参考。更多还原

【Abstract】 All kinds of unconventional emergencies with serious impacts occasionally happen in the world, including earthquakes, big terrorism attacks and other events that always result in a large number of casualties. As a special kind of relief material, the important role blood plays in emergency rescue. To cope with the sharp rise in demand for blood in emergency, emergency transferring of blood from unaffected area is required when blood inventory and donations in disaster area can not satisfy the clinical demands.Blood relief is directly related to the lives of the injured people, so it is of great social significance to research on emergency blood transferring problems in unconventional emergencies. During emergency blood relief, rational decision shall be made for the selection blood banks in emergency service, and how many blood quantities distributed to each affected blood bank shall be determined; multimodal transportation routes shall be optimized; and timely transshipment of overstocked blood in affected blood banks shall be applied so as to reduce of the risk of outdate; when shortage of blood is inevitable during blood distribution, the substitution relations among each blood type shall be considered to reach an optimization distribution plan of higher satisfaction.Main research contents in the paper are as follows:(1) By analyzing the blood characteristics and phased characteristic of emergency blood support, as well as considering the impact of transportation mode on blood quality, a multi-stage optimization model of selection for blood banks in emergency service and allocation for emergency blood was developed for emergency blood transferring problem with multi-type blood products and multiple affected areas. A genetic algorithm with vector coding based on giving priority to key affected areas was proposed, with its time complexity analyzed. Finally, the model and algorithm were verified by a numerical example based on the emergency blood supply process in Wenchuan earthquake. The results show the high efficiency of proposed genetic algorithm and the superiority of proposed multi-stage decision-making method in indexes such as the freshness and overstock of blood.(2) Considering the physiological and emergency supply characteristics of blood, as well as multi-type blood and multi-modal transportation mode, a combined optimization model for blood banks in emergency service selection and transportation routing problem in emergency blood supply was developed based on a two-phase decision-making approach. The aim is to minimize the last arrival time, maximize the lowest average freshness when received, and minimize the total transportation cost of emergency blood. A hybrid genetic-tabu algorithm with local neighborhood search was proposed. The model and algorithm were verified by a numerical example with different scenarios of blood demand, which are built based on the emergency blood supply process in’5.12’Wenchuan earthquake. The results show that weight coefficients of the objective function and the demand forecasting error will affect the blood transferring solution, multi-modal transportation could increase the timeliness and efficiency of blood transferring. And the robustness of two-phase decision-making approach was confirmed by the comparison with one-phase decision-making approach on the indexes such as the last arrival time, the average freshness, the total transportation cost, and the overstock of blood.(3) By analyzing blood physiological characteristics, substitution relations and priority level among blood groups were determined. A bi-level programming model was developed for emergency blood distribution with multi-group blood products in multiple affected areas by considering substitution strategy. The upper model is to minimize the blood shortage as the goal, and the lower model is to minimize blood group substitution weight. On the premise of the consistency of the optimal solutions, the model was transformed into a mixed integer programming problem with a single objective. After ratiocinating the substitution nature of blood, a multi-phase greed heuristic algorithm was proposed with its time complexity analyzed. Finally, the model and algorithm were verified by a numerical example based on the emergency blood distribution process after Wenchuan earthquake. The influence of substitution rate on distribution scheme was analyzed, and the blood demand satisfaction improved by substitution strategy was proved. The results of different scale numerical examples show that the heuristic algorithm has high adaptability and calculation efficiency by comparing with CPLEX software.(4) By analyzing the blood inventory status and outdating characteristic, the age-based properties of transshipping and receiving was deduced, and a transshipment model was set up to minimize transport costs. Finally, effectiveness of reasoning model in the paper, difference between transfer strategies as well as support reliability coefficient, demand saturation coefficient, decision-making time and blood shelf life’s impacts on transshipment decision-making was confirmed through example simulation. All these results provided blood transshipment decision-making in emergency support with theoretical basis and referential examples.更多还原