Using Memory in the Right Way to Accelerate Big Data Processing

Big data processing is becoming a standout part of data center computation. However, latest research has indicated that big data workloads cannot make full use of modern memory systems. We find that th...     

介观模拟研究温度对DPPC磷脂双层膜结构的影响

生物膜作为生物体的一种基本组成单元,在各种生命活动中占据重要地位。生物膜的结构和功能的稳定性是其发挥一切生理功能的前提。本文以DPPC磷脂双分子层为生物膜模型,采用基于Martini力场的粗粒化模拟方法,考察了温度对DPPC磷脂双层膜结构、形态和功能的影响。结果表明,315 K为DPPC磷脂双层膜的相转变温度;温度升高,DPPC分子间隙增加,尾链发生折叠弯曲,引起脂均面积增大,厚度减小,流动性增加;DPPC双层膜中甘油酯基团组成的栅栏区为膜中分子排列最有序的区域,也是阻碍小分子透膜的主要屏障区。研究结果不仅有助于更好地阐释生物膜的结构形态和功能之间的关系,也有助于理解药物与膜的相互作用,指导临床用药。     

1秒采集的三维头像摄影系统

为了实现高精度、快速度、低成本的三维头像建模,提出一个人体头部的三维摄影系统.该系统能够1 s采集用户的三维信息,3 min内建立照片真实感的三维模型.通过配置多目RGB-Depth相机摄影系统,保证1 s内完成对人体头部的全方位数据采集,进而对所得到的点云数据进行数据增强、配准、表面重建和着色,最终生成高逼真的三维头像模型.     

工业软件现场测试中的拆分及其测试数据设计

在工业软件的用户生产现场测试中,可能由于操作风险、用户生产限制等约束而导致测试不充分,针对实践中的难点提出新的现场测试过程及其测试数据生成方法。定义了测试拆分子集及相关属性概念,根据现场因素设计的测试拆分子集提高了现场测试可主动性;综合现场因素设计拆分方案,采用针对性的辅助管理程序引导每个测试拆分子集的现场测试的实施;通过对生产现场的历史数据的处理建立针对各测试拆分子集的测试准备数据库,以搜索测试准备数据库的结果为依据并结合现场因素设计现场主动测试的输入数据,并说明了实践中的综合实施过程。通过石化生产优化控制系统的现场测试的实际案例,说明了这种方法能够在预防风险的前提下显著提高现场测试的充分性和测试效率。     

多媒体教学中的屏幕分块算法的改进

随着多媒体技术的发展与普及,多媒体视频教学成为在线教学的一种趋势,这些应用中很重要的一个功能就是屏幕信息的捕获与传输。所以,如何以最短的时间、最少的带宽占用获取高质量的屏幕数据越来越成为研究热点。介绍了多媒体教学系统中采用的屏幕分块原理,提出了一种对客户端屏幕位图进行分块的算法,具体分块数采取动态规划思想获取,然后,对相邻时刻对应分块位图进行比较,最终只传输发生变化的分块到服务器以减少数据传输量,提高局域网屏幕监控的性能。     

An Improved Neural Network with Random Weights Using Backtracking Search Algorithm

This paper proposes a hybrid algorithm by combining backtracking search algorithm (BSA) and a neural network with random weights (NNRWs), called BSA-NNRWs-N. BSA is utilized to optimize the hidden layer parameters of the single layer feed-forward network (SLFN) and NNRWs is used to derive the output layer weights. In addition, to avoid over-fitting on the validation set, a new cost function is proposed to replace the root mean square error (RMSE). In the new cost function, a constraint is added by considering RMSE on both training and validation sets. Experiments on classification and regression data sets show promising performance of the proposed BSA-NNRWs-N.     

Experimental study of stable surfaces for anti-slug control in multi-phase flow

Severe slugging flow is always challenging in oil & gas production, especially for the current offshore based production. The slugging flow can cause a lot of problems, such as those relevant to production safety, fatigue as well as capability. As one typical phenomenon in multi-phase flow dynamics, the slug can be avoided or eliminated by proper facility design or control of operational conditions. Based on a testing facility which can emulate a pipeline-riser or a gas-lifted production well in a scaled-down manner, this paper experimentally studies the correlations of key operational parameters with severe slugging flows. These correlations are reflected through an obtained stable surface in the parameter space, which is a natural extension of the bifurcation plot. The maximal production opportunity without compromising the stability is also studied. Relevant studies have already showed that the capability, performance and efficiency of anti-slug control can be dramatically improved if these stable surfaces can be experimentally determined beforehand. The paper concludes that obtaining the stable surface on the new developed map can significantly improve the production rate in a control scheme. Even though the production rate can be further improved by moving the stable surface using advanced control strategies, the constant inputs can in some cases be preferable due to the easier implementation.     

k-Multi-preference query over road networks

Nowadays, the road network has gained more and more attention in the research area of databases. Existing works mainly focus on standalone queries, such as k-nearest neighbor queries over a single type of objects (e.g., facility like restaurant or hotel). In this paper, we propose a k-multi-preference (kMP) query over road networks, involving complex query predicates and multiple facilities. In particular, given a query graph, a kMP query retrieves of the top-k groups of vertices (of k facility types) satisfying the label constraints and their aggregate distances are the smallest. A naïve solution to this problem is to enumerate all combinations of vertices with k possible facility types and then select the one with the minimum sum distance. This method, however, incurs rather high computation cost due to exponential possible combinations. In addition, the existing solutions to other standalone queries are for a single type of facilities and cannot be directly used to answer kMP queries. Therefore, in this paper, we propose an efficient approach to process a kMP query, which utilizes an index with bounded space and reduces the computation cost of the shortest path queries. We also design effective pruning techniques to filter out false alarms. Through our extensive experiments, we demonstrate the efficiency and effectiveness of our proposed solutions.