模型算法控制器降阶方法的研究

使用模型算法控制理论设计的控制器,阶数较高,不利于工程实现和参数调整.针对这个问题,通过非参数模型和参数模型的转换对其进行降阶优化,调整闭环系统新极点的位置,可以获得动态特性和鲁棒性。仿真结果表明了该方法的有效性.     

油船货油保温时间计算

采用理论分析和实船应用检验相结合的方法,对大连远洋运输公司的72000 DWT双壳体原油船航行过程中,油舱货油自然冷却过程进行分析研究.根据非稳态导热集总参数法建立油船油舱保温数学模型,研制了油船货油保温计算软件.通过实船应用比较,软件计算的保温时间和实际的保温时间最大误差为13%.     

含区间参数的战场资源动态调度模型及算法

战场资源调度是作战指挥领域研究的热点.首先描述了战场资源调度问题,分析了战场资源动态调度需求,在此基础上建立了包含区间参数的战场资源动态调度模型.然后设计了三种不同的贪心策略,分别为基本贪心策略、双重贪心策略和一致贪心策略,提出了模型求解的混合贪心算法.最后结合联合作战算例进行了仿真验证,结果表明本文方法可行优越,能应用于不确定性的战场环境中.     

Species of the genera Bakera and Mangganeura (Homoptera: Auchenorrhyncha: Typhlocybinae) occurring on mango and other crop plants in the Oriental region

The genera Bakera (Bakera), Bakera (Guinobata) and Mangganeura (Homoptera: Auchenorrhyncha: Typhlocybinae) are re-described and compared. Sandalla is considered a synonym of Bakera (Guinobata) (syn. nov.).

Bakera (Bakera) nigrobilineata (Melichar) (=vittata Mahmood) is redescribed and illustrated from mango in the Philippines, on which it is frequently abundant, and found in association with Bakera (Bakera) luzonica sp. nov. which is described. Bakera (Bakera) abaca sp. nov. is described from the Philippines having been collected on abaca (Manila hemp). Bakera (Guinobata) nigroscuta Mahmood is redescribed from the Philippines and its association with coconut confirmed. Mangganeura reticulata Ghauri, described from Malaysia on mango, has been found in Thailand and the hitherto unknown female is described.     

Taxonomic revision of Ascidiacea (Tunicata) from the upper continental slope off north-western Australia

A collection of ascidians from 100–1000 m deep off the north-western coast of Australia has yielded 19 species, including one new, two possibly endemic to the north-western coast of Australia, four well-known Indo-west Pacific tropical species and 12 species known from the deeper waters of the western Pacific, of which 11 had not previously been recorded from off the western Australian coast or from other parts of the Indian Ocean. A third of these 12 species displays adaptations for a reduction in filter-feeding capacity, as do other known species from deeper slope and abyssal habitats. Nevertheless a connection with fauna of deep ocean basins of the Indian Ocean is not apparent at this stage. The full range of characters associated with species, genera and families are discussed, and those contributing to the identification of immature or mutilated specimens, as well as those relevant to an assessment of phylogenetic relationships, are emphasized.     

摩托车声品质评价的神经网络方法

针对声品质评价过程中线性回归模型评价结果的不足,采用BP神经网络对人的主观评价结果进行预测.采集摩托车在不同发动机转速下驾驶员耳旁的声信号样本,采用分组成对比较法进行主观评价试验,选取了响度、尖锐度、粗糙度作为神经网络模型输入参数,结合主观评价结果对模型进行训练与检验,并与线性回归模型输出结果进行比较.结果表明,选取驾驶员双耳响度、尖锐度、粗糙度作为模型输入能够较为准确地反映人耳对摩托车噪声的主观感觉.     

准同期参数测量算法研究

在发电机并网过程中,对信号的频率、相位、幅值等参数要进行快速、准确的测量,传统的测量方法参数误差比较大.本文提出利用窗函数及频谱校正的DFT算法来对信号进行参数测量,仿真结果表明,该算法不需要对电压信号进行整周期采样,可以对频率在较大范围内变化的信号进行准确测量,从而较好的解决了由于并列双方电压的频率不断变化而引起的测量误差大的问题.     

基于混合先验分布的贝叶斯因子分析模型

针对现有因子分析模型不能充分融合模型参数信息问题,通过研究因子分析模型的统计结构,构造了参数的混合先验分布;利用贝叶斯定理证明了模型因子载荷阵的条件后验分布为矩阵t分布,协方差阵的条件后验分布为逆Wishart分布.实证研究表明:由于参数先验分布的作用,贝叶斯因子分析结果与传统的因子分析之间存在明显的差异.     

网络攻防训练效果评估研究初探

从网络攻防技能训练效果评估的特点入手,分析了它与其它技能训练效果评估的相同点与不同点,提出了一种专门针对网络攻防技能的训练效果评估的研究方法。该方法较全面地考虑了影响效果评估的各种因素,提出了相应的解决办法。