数据仓库技术在决策支持系统的应用

简述了数据仓库,联机分析处理,数据挖掘的概念和技术,提出和实现了一种利用数据仓库技术及其工具,结合传统ＤＳＳ 的四库结构,设计和实现决策支持系统的新方法     

一种基于有效修剪的最大频繁项集挖掘算法

对关联挖掘中的最大频繁项集挖掘问题进行了研究,提出了一种基于项集格修剪机制的最大频繁项集挖掘算法.采用项集格生成树的数据结构,将最大频繁项集挖掘过程转化为对项集格生成树进行深度优先搜索获取所有最大频繁节点的过程. 其中提高算法效率的一个重要措施是在遍历项集格生成树的过程中对生成树进行修剪.给出了项集格生成树的三个性质,并在此基础上提出了直接超集修剪、间接超集修剪与事务集等价修剪三种修剪机制,尽可能忽略非频繁节点及其所生成的扩展节点以减少遍历的节点数目.试验结果表明,三种修剪机制都能够有效地减少搜索空间,其中事务集等价修剪机制的效果最好,算法的性能与输入数据集的稠密程度相关.      

基于数据挖掘技术的液体火箭发动机试验数据分析和处理

分析了数据挖掘技术的研究概况,研究了偏差检测、人工神经网络等数据挖掘方法,实现了挖掘方法在液体火箭发动机试验数据分析处理中的应用。最后,用某发动机试验数据进行了验证。     

基于Terminal滑模的小行星探测器着陆连续控制

针对弱引力场不规则小行星探测器安全下降与着陆,提出了一种基于比例微分(PD,Proportional plus Derivative)及非奇异Terminal滑模的连续控制方法.在着陆点坐标系下推导了探测器的动力学方程,设计了开环燃料次最优多项式标称轨迹制导方法.针对下降和着陆两个不同阶段提出了PD和非奇异Terminal滑模变结构连续控制方法.将PD控制的易操作性与非奇异Terminal滑模控制收敛速度快、调整时间短有效结合,保证了探测器安全着陆.仿真结果表明了提出方法的可行性和有效性.     

双体小行星系统平衡态与稳定性研究

由于双体小行星独特的运动形式可为行星的演化提供重要线索,因而成为小行星探测中的热点目标。基于双椭球体模型研究了双小行星系统的相对运动、平衡态及稳定性。首先基于双椭球的全二体模型建立了双星系统相对运动的动力学方程;其次利用拉格朗日方程,通过求解系统角动量和能量,建立了双星系统平衡态对应的状态约束;最后给出了通过零速度状态曲面判断双星系统平衡态稳定性的一般性方法,在此基础上分析了小行星物理参数变化对系统平衡态稳定性的影响。研究可为未来双体小行星系统探测任务中的轨道设计与控制提供重要的理论参考。     

Landing of the asteroid probe with three-legged cushioning: Planar asymmetric dynamics and safety margin

The probes landing on the surfaces of the asteroids can increase the scientific return of the exploration missions and also promote the development of deep space resources. Because of its excellent applicability to the uneven terrain and a lighter configuration than the four-legged mechanisms, the three-legged cushioning mechanisms are suitable for dissipating the impact energy and then quickly stabilizing the probe attitude when the probe lands on the micro-gravitational surfaces of the asteroids. Research on the landing dynamics of the probe facilitates to the design of the landing-cushioning mechanism and the optimization of its configuration, as well as the assessment of the landing safety. Comparing with the previous extensive related literature focusing on landing dynamics of the probes assisted by the four-legged cushioning mechanisms, this paper studies creatively the planar dynamics considering the asymmetric characteristic and the leg-leg coupling to understand the landing process of the asteroid probe with the three-legged cushioning mechanism and thereby to optimize the configuration of cushioning mechanism and assess safety margin of the landing. According to the touchdown status, the asymmetric landing modes are classified and the coupling issue in the construction of the landing models is explained. Consequently, two types of dynamics models describing the two-stages touchdown cushioning process of the probe are established. Then, five significant configuration factors of the cushioning mechanism are extracted, and their values combinations are designed according to the Taguchi orthogonal method. On this basis, the maximum safe landing attitude angles of the probe are solved by using these values combinations as the input conditions under the dangerous situations in different landing modes. The range analysis and nonlinear fitting methods are employed to discuss the influence of the configuration factors on the landing safety margin, and the favorable parameter values of the configuration factors are determined. Next, the influence of the ground obstacle on the landing safety margin and several methods to improve the margin are researched. Finally, the complete attitude changes of the probe in two representative landing cases are analyzed. The results studied in this paper can contribute to configuration optimization of the three-legged cushioning mechanisms and safety assessment of the legged probes landing on the asteroids, as well as to provide a reference for discussing the leg-leg coupling issue received less attention in landing dynamics of the probes with the four-legged cushioning mechanisms.     

基于平衡核函数聚类的飞行航迹数据分析方法

目前,使用数据挖掘的方法对目标的飞行航迹进行分析来确定航迹类别具有许多应用价值。飞行航迹数据具有维数高、交连多、可分类性能差等特点,要做到尽可能精确的聚类和分类十分困难。文章立足提高飞行航迹数据聚类分析的准确性,在航迹特征数据的预处理阶段,提出了一种平衡核函数的K-均值聚类方法,可以解决高维特征数据带来的奇异性,还能提高交叠样本的聚类性能;设计了一种模糊支持向量机的算法框架实现航迹的分类。通过实际飞行航迹数据集测试了设计框架下航迹聚类和分类识别的有效性,在实际工程上具有广泛的应用前景。     

基于数据挖掘和小波神经网络的航材消耗预测方法

运用数据挖掘技术对航材消耗的历史数据进行关联分析，筛选小对保障飞机飞行有重要作用的航材消耗数据，大大缩减了需要预测的航材数量，同时对消耗航材之间的内在影响关系进行量化。在分析人工鱼群算法原理的基础上，对算法中步长参数和视野范同参数的设置方法进行了改进。实例结果表明，运用小波神经网络预测航材消耗的方法大大降低了预测误差，说明了该方法的有效性、可行性和实用性。