基于BP神经网络的D-S证据理论及其应用

命题基本概率分配(BPA)的确定是D-S证据理论得以广泛应用的关键之一.目前,大部分确定方法受专家知识偏好影响较大,难以反映客观情况.将BP网络运用到基本概率分配的确定过程中,使得BP网络和D-S证据理论两者有机地联合应用,这样既可利用D-S证据理论来表达和处理不确定信息,又可以充分发挥BP网络的自学习、自适应和容错能力.文中建立了基于BP网络的D-S证据理论的故障诊断模型,并给出了证据的融合算法.仿真实验表明,该模型可行.     

基于多元高斯概率分布的无线局域网定位方法研究

针对无线局域网环境及室内目标定位需求,研究了基于接收信号强度指示的定位技术,提出了一种基于多元高斯概率分布的信号强度模式匹配方法,讨论了定位系统框架。试验结果表明,该方法具有较好的室内定位效果。     

大区域防空的进化群决策机制研究

大区域防空所涉及的区域广阔,因素众多,其作战计划的制定比较复杂。为解决此问题,提出了一种基于改进德尔菲法的具有进化学习功能的群决策机制,得到了纳什均衡解,使作战方案的制定更加合理,有助于防空作战效能的提高。     

基于Dempster-Shafer证据理论的特征融合在军用电子设备模块化故障诊断过程中的应用

军用电子设备的维修是以更换电子集成模块为主,根据这一特点,通过分析处理电子设备的故障特征信息,把一种基于D-S证据理论的算法应用于故障特征信息的融合,从而实现故障电子模块的定位.结合实例,论证了这一方法在军用电子设备模块化故障诊断中的可行性.     

基于B样条的螺旋桨升力面设计

为了改善螺旋桨的空泡性能,从效率和空泡性能2个方面的权衡来确定环量分布.在升力面理论设计中计及桨毂的影响.桨叶几何用较少的B样条控制角点来表示.将升力面理论计算的控制点处法向速度的平方和作为优化目标函数,并通过最小化目标函数得到桨叶几何.     

Complexity and workload considerations in product mix decisions under the theory of constraints

The literature on the product mix decision (or master production scheduling) under the Theory of Constraints (TOC), which was developed in the past two decades, has addressed this problem as a static operational decision. Consequently, the developed solution techniques do not consider the system's dynamism and the associated challenges arising from the complexity of operations during the implementation of master production schedules. This paper aims to address this gap by developing a new heuristic approach for master production scheduling under the TOC philosophy that considers the main operational factors that influence actual throughput after implementation of the detailed schedule. We examine the validity of the proposed heuristic by comparison to Integer Linear Programming and two heuristics in a wide range of scenarios using simulation modelling. Statistical analyses indicate that the new algorithm leads to significantly enhanced performance during implementation for problems with setup times. The findings show that the bottleneck identification approach in current methods in the TOC literature is not effective and accurate for complex operations in real‐world job shop systems. This study contributes to the literature on master production scheduling and product mix decisions by enhancing the likelihood of achieving anticipated throughput during the implementation of the detailed schedule. © 2015 Wiley Periodicals, Inc. Naval Research Logistics 62: 357–369, 2015     

Synchronization in hub terminals with the circular arrangement problem

Hub terminals are important entities in modern distribution networks and exist for any transportation device, that is, cross docks and parcel distribution centers for trucks, container ports for ships, railway yards for trains, and hub airports for aircraft. In any of these hubs, the mid‐term planning task of synchronizing the transshipment of goods and passengers when servicing the transportation devices has to be solved, for which many different solution approaches specifically tailored to the respective application exist. We, however, take a unified view on synchronization in hubs and aim at a general building block. As a point of origin, a basic vertex ordering problem, the circular arrangement problem (CAP), is identified. We explain the relation between the CAP and hub processes, develop suited algorithms for solving the CAP, and extend the basic CAP by multiple additions, for example, arrival times, limited storage space, and multiple service points, make the problem adaptable to a wide range of hub terminals. © 2015 Wiley Periodicals, Inc. Naval Research Logistics 62: 454–469, 2015     

北斗静止轨道卫星信号盲区快速并行解算方法

北斗系统静止轨道卫星信号盲区解算方法复杂,串行计算耗费时间长,须在并行环境下利用更多的计算资源进行北斗盲区的快速解算。在分析北斗盲区解算原理与算法并行特征基础上,提出动态盲区影响域的并行解算方法。以栅格单元为并行粒度进行任务划分,实现了北斗盲区的高效并行解算。运用全国范围59景数字高程模型数据,利用8进程进行盲区并行解算,耗费时间约为5小时。实验测试结果表明:算法的并行效率随着进程数的增加有所衰减,但稳定在96%以上。其程序中间件已集成应用于高性能地理信息平台,应用效果良好。     

训练电磁环境生成和逼真度评估方法

构建与战场电磁环境尽可能相似的训练电磁环境、评估所构建的训练电磁环境与战场电磁环境的相似程度,是基地训练电磁环境构建必须解决的突出问题。基于相似理论基本原理和相似度计算基本方法,结合基地训练电磁环境构建实践,采用精确值相似度计算生成训练电磁环境,采用区间值相似度计算评估训练电磁环境的逼真度。     

军事情报价值的当量评估法

研究了军事情报价值的综合性和相对性,提出了军事情报价值评估的当量分析法,分别针对确定性标度和模糊标度结合层次分析法给出了评估算法。算例结果表明方法有一定的合理性且评价结论易于理解。