【作者】 王文峰；

【导师】 郭波；

【作者基本信息】 国防科学技术大学 ， 管理科学与工程， 2008， 博士

【摘要】 现代高技术战争中装备保障的重要性愈加突出,对保障系统的反应速度和保障能力要求越来越高。装备保障网络是以装备保障设施为依托,将各种保障资源按照一定的要求和原则合理部署而最终形成的一个网络化布局的保障体系,其优化设计是装备保障系统构建决策过程中的重要工作和关键环节,最终的决策方案对保障系统的服务能力、服务质量、运行成本都具有直接、重要的影响。论文结合装备保障问题的特殊性,借鉴现代供应链与物流理论、决策理论和规划方法等的研究成果,对装备保障网络优化设计问题进行深入研究,目的是结合不同的决策时机、决策目标以及不同的保障业务类型,探讨装备保障网络的优化设计方法。主要研究内容如下:(1)针对装备保障网络设计问题的多利益主体特性,提出了综合多属性决策的保障网络设计方法。该方法首先从安全性条件、交通条件、展开条件和水电能源条件以及气候、地质结构条件等方面抽取决策属性,建立评价指标体系,由装备保障部门据此对保障设施备选位置进行多准则评价,确定可用的设施备选址点集合;然后,综合考虑被保障单位对保障服务的实际要求(如对物资的品种、数量、以及保障时效性要求等)与保障部门的实际保障能力,建立网络设计决策的数学规划模型,以在备选点集合中选择建设保障设施,确立保障服务关系,并对应配置资源。该方法实现了多属性决策方法和数学规划方法的优势互补。(2)建立了装备战略保障设施的多级覆盖设施选址模型,提出了基于分布估计算法的模型求解算法。首先结合战略保障设施的作用和建设时机,分析总结了战略保障所具有的多源保障和多重时效性标准特征;然后,在现有的最大覆盖设施选址模型基础上对“覆盖”概念进行了有效地扩展,以使有限的设施及其储备的资源对潜在作战方向或作战区域的保障效果最大化为目标,建立了装备战略保障多级覆盖设施选址模型,并基于分布估计算法研究了模型求解方法;最后,通过计算示例进一步阐述了多级覆盖建模的思想,并用实验分析了算法参数对算法质量和效率的影响。(3)建立了战役供应保障网络设计模型,设计了模型的启发式求解框架。以既定战役作战任务和作战方案条件下的装备弹药供应保障问题为背景,对装备保障前进基地和补给站的设施选址问题进行了研究,建立了综合考虑战役供应保障系统的多层次性、多物资特性、多种保障方式综合运用等特性,以及装备保障的时效性要求、保障设施的容量限制和运力限制等因素的网络优化设计模型,模型目标是最大化能及时运达的作战物资对战役任务的支撑程度。然后基于禁忌搜索算法研究建立了模型的启发式求解框架,并以计算示例验证了该模型和算法的有效性。(4)以防御作战样式下陆军装甲装备的战役维修保障为背景,研究了战役装备维修保障网络优化设计问题。首先分析了该问题的特性,然后根据维修资源有限与否、资源有限条件下的资源配置方法以及保障单位与作战单位之间保障关系的确立规则,分六种情形建立了不同的决策环境下由维修基地和维修站两级设施构成的维修保障网络设计问题的数学规划模型,给出了模型的启发式求解过程。模型优化的目标是在保证装备维修保障的军事性目标,即损伤装备在各维修设施处的无等待修复率高于给定水平的前提下,优化保障设施的选址、设施与作战单位间的保障关系以及设施处的资源配置方案,以最小化系统的总期望成本,提高保障网络建设的经济性。最后用计算示例对所建立的优化模型和求解过程进行了验证。文章最后总结了全文的研究内容,并对进一步的研究工作进行了展望。更多还原

【Abstract】 The prominent importance of equipment support in the high-tech war puts forward the higher requirement for the response velocity and support capability of the support system. The equipment support network is the basis to provide support service. All kinds of resources are deployed on the facilities in the network such as military bases and warehouses to constitute a support system according to rules and requirements. The design of the support network is an important aspect to construct an enquipments support system, and the decision has direct and crucial effect on the service ability, the quality of support, and the cost of the support system.Taking account of the particularity of equipment support problem, this thesis researches the optimal design problem of equipment support network based on the investigation results of supply chain and logistics, decision science and mathematics programming. It aims to find some methods to optimal design the support network in consideration of different decision opportunities, decision objects and different support operation types. The following are the major contributions of this dissertation:Firstly, the multi-attribute integrated support network design method is proposed, which considers the multi-stockholders characteristic of the problem. According to the method, the equipment support departments extract some attributes such as safety conditions, traffic conditions, deployment conditions, energy-obtaining ability, climate and geologic conditions to construct the indices. And the practicable sites can be found from the candidates by a multicriteria decision progress with the indices. Then the network optimal design mathematical programming models would be presented to locate the facilities, establish the service relationships and allocate resources. In the models the requirements for the support services of different supported units and the support ability of the support department are considered. This method takes the advantages of the multi-attribute decision method and of the planning method to reinforce each other.Secondly, the multi-quality-of-coverage facility location model for the strategic support facilities is presented, and its solution approach is provided. The multi sources and multi time limits characteristics of strategic support are analyzed according to the strategic support facilities’function and construction opportunity. Then the definition of‘covering’is extended based on the maximal covering location problem, and a multiple quantity-of-coverage facility location model is proposed to maximize the support effect of limited number of strategic support facilities to the potential conflict areas. In the model the multiple sources need of a demand point and the different time requirements to the sources when the services are from many facilities are typically considered. A heuristic based on Estimation of Distribution Algorithms is developed to solve the problem. Moreover, an illustrative example is given to explicate on the model, and the influences of the parameters to the algorithm’s quality and efficiency are given based on experiments.Thirdly, the combat logistics network optimal design model and its solution framework are investigated. The characteristics of the depots’location problem are analyzed with the background of ammunition supply problem in a compaign with the given mission and the combat project. A multi-commodity two-stage capacitated facility location-transportation model is proposed to maximize the support of the resources arrived timely to the combat. In the model the capacity limitation of the facilities, the transportation capability at the facilities, the support time requirements for different resources at different need points, and the multi-modal network flows have been taken into account. Then a solution approach based on Tabu Search algorithm is given, and the results demonstrate the validity of the model.Fourthly, the optimal design problem of cavalryman armaments repair network with 2-level repair facilities of repair depots and the repair base under the defence combat pattern is considered. The characteristics of the problem are discussed. According to whether the servers are limited or not, whether the server distribution plan could be adjusted or not, and whether the service relationships between combat units and repair facilities have been assigned in advance or not, it’s devided into 6 sub-problems. Then the integrated optimization models are proposed to find locations for the repair depots, to find the best service relationships between combat units and the depots, and to decide the number of servers and base stock levels which should be allocated at the repair facilities for different decision conditions, so that the total expected cost to construct and operate the support system is minimized and the probabilities that damaged armaments could be repaired by the facilities without waiting are above some specified levels. At the same time the heuristics for the models are proposed. The experimental results verify the models and algorithms.Finally, remarks are made for all researches and some plans for future work are presented.更多还原