大型注水系统注水站合理排量的确定方法

大型注水管网系统结构复杂 ,涉及的井点、管网节点多 ,且压力、流量互相耦合 ,注水站排量合理分配难度大。以流体力学基本理论为基础 ,利用“压力谷”的性质 ,根据注水管网特性 ,建立了注水管网的基本模型 ,并在满足注水量和管路阻力损失最小的前提下 ,给出了一种确定大型注水系统注水站设计排量和运行排量的方法。实例计算表明 ,利用该方法调节大型注水系统注水站排量 ,可以使注水站和管网同时达到最优运行状态 ,并且能够科学地计算注水站的设计排量。该研究为大型注水系统的合理设计和优化运行提供了一定的理论指导     

The cluster-galaxy cross-correlation and the average infall velocity around cluster in the modified CDM models

The cluster-galaxy cross-correlation and the average infall velocity around cluster for the tilted CDM models, the hot-cold mixed dark matter models (MDM) and theA CDM models are calculated. And it is found that the prediction of the cluster-galaxy cross-correlation of theA CDM models with Ω_A=0.9, Ω_C=0.1,h=1,b = 1 or 2 is much higher than that of observation for the radius larger than 10 h⁻¹ Mpc (H₀ = 100 h.kms⁻¹Mpc⁻¹), and the other models are compatible with the observation on the scale of (3–20) h⁻ Mpc. The results can be used to compare with new observations to restrict the nature of the dark matter and the biasing factor. Project supported by the National Natural Science Foundation of China (Grant No. 19733001).     

Designing information structures for quadratic decision problems

This paper investigates the problem of unconstrained optimization when there is only partial information on the random parameters in the objective function. The relation between the optimization performance and the available information is established. The best information structure design with fixed rank is described. The designing procedure is set up in such a way that successive information augmentation or deletion can be considered. The procedure can also be extended to multiperson decision problems.     

用楔角偏差改善光楔列阵系统焦线的大尺度不均匀性

在加工光楔列阵线聚焦系统时，人为地使各单元楔角与精确的理论值之间存在一定的偏差，以使单元衍射引起的大尺度不均匀性得到改善，文中进行了理论计算和实验验证，在既能改善均匀性，又能保证一定的焦线能量利用率的情况下，给出了合理的偏差数据，为优化均匀聚焦系统的工程设计提供了必要的依据。     

Lp-based combinatorial problem solving

A tutorial outline of the polyhedral theory that underlies linear programming (LP)-based combinatorial problem solving is given. Design aspects of a combinatorial problem solver are discussed in general terms. Three computational studies in combinatorial problem solving using the polyhedral theory developed in the past fifteen years are surveyed: one addresses the symmetric traveling salesman problem, another the optimal triangulation of input/output matrices, and the third the optimization of large-scale zero-one linear programming problems.     

丘陵地区高速公路景观设计研究

以G85成渝高速公路（永川段）为例,探讨丘陵地区高速公路生态用地的帚观规划以及建设模式．在剖析公路景观中4个典型区域的规划模式及其植被配置方式的基础上,提出大尺度景观的设计理念：根据公路途经地域的不同,设计田园乡村绿化景观、山地丘陵绿化景观和村镇绿化景观,以展示沿途的田园风光、山地风景和村镇绿化景观;结合有色植物的合理配置,展示景观动态．     

基于改进蚁群算法的大型综合建筑火灾疏散路径规划

动态逃生指示系统主要应用于大型综合建筑物内,此系统可根据建筑物内发生的火灾等突发情况动态指示人员疏散逃生,缩短逃生时间提高逃生成功率。通过研究现有动态逃生指示系统路径规划问题,提出一种改进蚁群算法,将Dijkstra算法和蚁群算法相结合,利用Dijkstra算法的全局搜索能力,调整了蚁群算法启发函数中初始信息素分布情况,同时结合探测到的火灾实时信息对蚁群算法的启发函数,转移概率,信息素挥发系数和更新规则进行改进。通过仿真实验表明改进的蚁群算法提高了搜索效率和全局搜索能力,降低了陷入局部最优的可能性并优化了逃生路线。     

干扰状态下大规模网络数据流的分发协调技术研究

干扰状态下大规模网络数据流具有实时性、波动性、突发性等特点,当前数据流分发协调方法无法保证数据流的平稳传输及分发协调的实时性。为此,提出一种新的干扰状态下大规模网络数据流分发协调方法,通过网络数据流预测值和实际值之间的误差滤除干扰状态下的干扰数据,保证剩余数据流的纯正性。将大规模网络数据流分发协调问题转换成多目标优化问题,以提高大规模网络数据流分发协调的整体性能为目的设定一个目标函数,提出一种大规模网络数据流分发协调预测技术,通过比较处理器计算资源和未来的预期负载确定数据流分发协调策略。实验结果表明,采用所提方法对干扰状态下大规模网络数据流进行分发协调,不仅效率高、实时性好,而且CPU资源利用率很高。     

Transition region where the large-scale and very large scale motions coexist in atmospheric surface layer: wind tunnel investigation

The turbulent structures in atmospheric surface layer (ASL) are investigated in wind tunnel with hot-wire anemometers in present study. The results show that there exist two typical turbulent structures characterised by their streamwise length scales, i.e. large-scale motions (LSMs) and very large scale motions (VLSMs) as reported recently in pipe flow, channel flow, zero-pressure-gradient turbulent boundary layers and near-neutral ASLs. Moreover, a transition region containing both LSMs and VLSMs is first reported in present investigation, and this region depends on the Reynolds numbers. In the transition region, as the wall-normal distance increases, the turbulent energy ratio of LSMs is gradually weakened but VSLMs strengthened.     

基于Relief的组合式特征选择

ReliefF是公认的效果较好的filter式特征评估方法,但该方法一大缺点是不能辨别冗余特征．提出两种基于Relief的组合式特征选择算法：ReCorre和ReSBSW,这两种算法均首先利用ReliefF算法过滤掉无关特征,然后分别采用相关分析(Correlation)以及顺序后向搜索(SBS)的Wrapper算法去除冗余特征．在实际数据集以及人造数据集上进行了实验,分析比较了Relief,ReCorre以及ReSBSW算法的性能．实验结果得出如下结论：ReliefF方法对无关特征较多的数据集能够很好的降维,但对于实际数据中特征间关系较复杂的情况,只能去掉很少的无关特征,并会去除一部分相关特征,ReliefF不能处理冗余特征,ReCorre可以在ReliefF基础上去除大部分冗余特征．ReSBSW算法可得到较好的泛化性能,但算法计算量很高,不适合大规模数据集．