- Chao Xue;Chuang Lin;Jie Hu;
Rapid advancement of distributed computing systems enables complex services in remote computing clusters. Massive applications with large-scale and disparate characteristics also create high requirements for computing systems. Cloud computing provides a series of novel approaches to meet new trends and demands.However, some scalability issues have to be addressed in the request scheduling process and few studies have been conducted to solve these problems. Thus, this study investigates the scalability of the request scheduling process in cloud computing. We provide a theoretical definition of the scalability of this process. By modeling the scheduling server as a stochastic preemptive priority queue, we conduct a comprehensive theoretical and numerical analysis of the scalability metric under different structures and various environment configurations. The comparison and conclusion are expected to shed light on the future design and deployment of the request scheduling process in cloud computing.
2019年03期 v.24 249-261页 [查看摘要][在线阅读][下载 3193K] [下载次数：70 ] |[网刊下载次数：0 ] |[引用频次：2 ] |[阅读次数：1 ] - Ke Lin;Shangkai Gao;Xiaorong Gao;
For Brain-Computer Interface(BCI) systems, improving the Information Transfer Rate(ITR) is a very critical issue. This study focuses on a Steady-State Visually Evoked Potential(SSVEP)-based BCI because of its advantage of high ITR. Unsupervised Canonical Correlation Analysis(CCA)-based method has been widely employed because of its high efficiency and easy implementation. In a recent study, an ensemble-CCA method based on individual training data was proposed and achieved an excellent performance with ITR of 267 bit/min.A 40-target SSVEP-BCI speller was investigated in this study, using an integration of Minimal-Distance(MD) and Maximal-Phase-locking value(MP) approaches. In the MD approach, a spatial filter is developed to minimize the distance between the training data and the reference sine signal, and in this study, two different types of distance were compared. In the MP approach, a spatial filter is developed to maximize the Phase-Locking Value(PLV)between the training calibration data and the reference sine signal. In addition to the fundamental frequency of stimulation, the harmonics were used to train MD and MP spatial filters, which formed spatial filter banks. The test data epoch was multiplied by the MP and MD spatial filter banks, and the distances and PLVs were extracted as features for recognition. Across 12 subjects with a 0.4 s-data length, the proposed method realized an average classification accuracy and ITR of 93% and 307 bit/min, respectively, which is significantly higher than the current state-of-the-art method. To the best of our knowledge, these results suggest that the proposed method has achieved the highest ITR in SSVEP-BCI studies.
2019年03期 v.24 262-270页 [查看摘要][在线阅读][下载 1556K] [下载次数：114 ] |[网刊下载次数：0 ] |[引用频次：14 ] |[阅读次数：1 ] - Luning Liu;Xin Chen;Zhaoming Lu;Luhan Wang;Xiangming Wen;
Under the situations of energy dilemma, energy Internet has become one of the most important technologies in international academic and industrial areas. However, massive small data from users, which are too scattered and unsuitable for compression, can easily exhaust computational resources and lower random access possibility, thereby reducing system performance. Moreover, electric substations are sensitive to transmission latency of user data, such as controlling information. However, the traditional energy Internet usually could not meet requirements. Integrating mobile-edge computing makes energy Internet convenient for data acquisition,processing, management, and accessing. In this paper, we propose a novel framework for energy Internet to improve random access possibility and reduce transmission latency. This framework utilizes the local area network to collect data from users and makes conducting data compression for energy Internet possible. Simulation results show that this architecture can enhance random access possibility by a large margin and reduce transmission latency without extra energy consumption overhead.
2019年03期 v.24 271-280页 [查看摘要][在线阅读][下载 3790K] [下载次数：123 ] |[网刊下载次数：0 ] |[引用频次：20 ] |[阅读次数：1 ] - Haowei Zhang;Junwei Xie;Jiaang Ge;Junpeng Shi;Zhaojian Zhang;
An efficient task-scheduling algorithm in the Digital Array Radar(DAR) is essential to ensure that it can handle a large number of requested tasks simultaneously. As a solution to this problem, in this paper, we propose an optimization model for scheduling DAR tasks using a hybrid approach. The optimization model considers the internal task structure and the DAR task-scheduling characteristic. The hybrid approach integrates a particle swarm optimization algorithm with a genetic algorithm and a heuristic task-interleaving algorithm. We introduce the chaos theory to optimize initialized particles and use entropy theory to indicate the diversity of particles and adaptively adjust the inertia weight, the crossover probability, and the mutation probability. Then, we improve both the efficiency and global exploration ability of the hybrid algorithm. In the framework of the swarm exploration algorithm, we include a heuristic task-interleaving scheduling algorithm, which not only utilizes the wait interval to transmit or receive subtasks, but also overlaps the receive intervals of different tasks. In a large-scale simulation,we demonstrate that the proposed algorithm is more robust and effective than existing algorithms.
2019年03期 v.24 281-290页 [查看摘要][在线阅读][下载 1881K] [下载次数：76 ] |[网刊下载次数：0 ] |[引用频次：22 ] |[阅读次数：1 ] - Xiaohu Yuan;Shaojun Guo;Chunwen Li;Bin Lu;Shuli Lou;
The centroid location of a near infrared star always deviates from the real center due to the effects of surrounding radiation. To determine a more accurate center of a near infrared star, this paper proposes a method to detect the star's saliency area and calculate the star's centroid via the pixels only in this area, which can greatly decrease the effect of the radiation. During saliency area detection, we calculated the boundary connectivity and gray similarity of every pixel to estimate how likely it was to be a background pixel. Aiming to simplify and speed up the calculation process, we divided the near infrared starry sky image into super pixel maps at multi-scale by Simple Linear Iterative Clustering(SLIC). Second, we detected the saliency map for every super pixel map of the image. Finally, we fused the saliency maps according to a weighted coefficient that is determined by the least square method. For the images used in our experiment, we set the multi-scale super pixel numbers to 100, 150,and 200. The results show that our method can obtain an offset variance of less than 0.27 for the center coordinates compared to the labelled centers.
2019年03期 v.24 291-300页 [查看摘要][在线阅读][下载 2894K] [下载次数：26 ] |[网刊下载次数：0 ] |[引用频次：3 ] |[阅读次数：1 ] - Zhiming Xu;Yu Bai;Kuangyu Liu;Shuning Wang;
This paper works on a heuristic algorithm with determinacy for the global optimization of Continuous PieceWise Linear(CPWL) programming. The widely applied CPWL programming can be equivalently transformed into D.C. programming and concave optimization over a polyhedron. Considering that the super-level sets of concave piecewise linear functions are polyhedra, we propose the Hill Tunneling via Weighted Simplex Centroid(HTWSC) algorithm, which can escape a local optimum to reach the other side of its contour surface by cutting across the super-level set. The searching path for hill tunneling is established via the weighted centroid of a constructed simplex. In the numerical experiments, different weighting methods are studied first, and the best is chosen for the proposed HTWSC algorithm. Then, the HTWSC algorithm is compared with the hill detouring method and the software CPLEX for the equivalent mixed integer programming, with results indicating its superior performance in terms of numerical efficiency and the global search capability.
2019年03期 v.24 301-316页 [查看摘要][在线阅读][下载 4888K] [下载次数：29 ] |[网刊下载次数：0 ] |[引用频次：0 ] |[阅读次数：1 ] - Amina Belhassena;Hongzhi Wang;
With the rapid development and wide use of Global Positioning System in technology tools, such as smart phones and touch pads, many people share their personal experience through their trajectories while visiting places of interest. Therefore, trajectory query processing has emerged in recent years to help users find their best trajectories. However, with the huge amount of trajectory points and text descriptions, such as the activities practiced by users at these points, organizing these data in the index becomes tedious. Therefore, the parallel method becomes indispensable. In this paper, we have investigated the problem of distributed trajectory query processing based on the distance and frequent activities. The query is specified by start and final points in the trajectory, the distance threshold, and a set of frequent activities involved in the point of interest of the trajectory.As a result, the query returns the shortest trajectory including the most frequent activities with high support and high confidence. To simplify the query processing, we have implemented the Distributed Mining Trajectory R-Tree index(DMTR-Tree). For this method, we initially managed the large trajectory dataset in distributed R-Tree indexes.Then, for each index, we applied the frequent itemset Apriori algorithm for each point to select the frequent activity set. For the faster computation of the above algorithms, we utilized the cluster computing framework of Apache Spark with MapReduce as the programing model. The experimental results show that the DMTR-Tree index and the query-processing algorithm are efficient and can achieve the scalability.
2019年03期 v.24 317-332页 [查看摘要][在线阅读][下载 2213K] [下载次数：94 ] |[网刊下载次数：0 ] |[引用频次：18 ] |[阅读次数：1 ] - Jingchun Cheng;Yali Li;Jilong Wang;Le Yu;Shengjin Wang;
Gender classification is an important task in automated face analysis. Most existing approaches for gender classification use only raw/aligned face images after face detection as input. These methods exhibit fair classification ability under constrained conditions, in which face images are acquired under similar illumination with similar poses. The performances of these methods may deteriorate when face images show drastic variances in poses and occlusion as routinely encountered in real-world data. The reduction in the performances of current gender classification methods may be attributed to the sensitiveness of features to image translations. This work proposes to alleviate this sensitivity by introducing a majority voting procedure that involves multiple face patches.Specifically, this work utilizes a deep learning method based on multiple large patches. Several Convolutional Neural Networks(CNN) are trained on individual, predefined patches that reflect various image resolutions and partial cropping. The decisions of each CNN are aggregated through majority voting to obtain the final gender classification accurately. Extensive experiments are conducted on four gender classification databases, including Labeled Face in-the-Wild(LFW), CelebA, ColorFeret, and All-Age Faces database, a novel database collected by our group. Each individual patch is evaluated, and complementary patches are selected for voting. We show that the classification accuracy of our method is comparable with that of state-of-the-art systems. This characteristic validates the effectiveness of our proposed method.
2019年03期 v.24 333-345页 [查看摘要][在线阅读][下载 1526K] [下载次数：55 ] |[网刊下载次数：0 ] |[引用频次：8 ] |[阅读次数：1 ] - Hongjie Fan;Zhiyi Ma;Hongqiang Li;Dongsheng Wang;Junfei Liu;
Community Question Answering(CQA) in web forums, as a classic forum for user communication,provides a large number of high-quality useful answers in comparison with traditional question answering.Development of methods to get good, honest answers according to user questions is a challenging task in natural language processing. Many answers are not associated with the actual problem or shift the subjects,and this usually occurs in relatively long answers. In this paper, we enhance answer selection in CQA using multidimensional feature combination and similarity order. We make full use of the information in answers to questions to determine the similarity between questions and answers, and use the text-based description of the answer to determine whether it is a reasonable one. Our work includes two subtasks:(a) classifying answers as good, bad, or potentially associated with a question, and(b) answering YES/NO based on a list of all answers to a question. The experimental results show that our approach is significantly more efficient than the baseline model, and its overall ranking is relatively high in comparison with that of other models.
2019年03期 v.24 346-359页 [查看摘要][在线阅读][下载 1831K] [下载次数：50 ] |[网刊下载次数：0 ] |[引用频次：10 ] |[阅读次数：1 ] - Wenchang Zhang;Fuchun Sun;Hang Wu;Chuanqi Tan;Yuzhen Ma;
The control of a high Degree of Freedom(DoF) robot to grasp a target in three-dimensional space using Brain-Computer Interface(BCI) remains a very difficult problem to solve. Design of synchronous BCI requires the user perform the brain activity task all the time according to the predefined paradigm; such a process is boring and fatiguing. Furthermore, the strategy of switching between robotic auto-control and BCI control is not very reliable because the accuracy of Motor Imagery(MI) pattern recognition rarely reaches 100%. In this paper, an asynchronous BCI shared control method is proposed for the high DoF robotic grasping task. The proposed method combines BCI control and automatic robotic control to simultaneously consider the robotic vision feedback and revise the unreasonable control commands. The user can easily mentally control the system and is only required to intervene and send brain commands to the automatic control system at the appropriate time according to the experience of the user. Two experiments are designed to validate our method: one aims to illustrate the accuracy of MI pattern recognition of our asynchronous BCI system; the other is the online practical experiment that controls the robot to grasp a target while avoiding an obstacle using the asynchronous BCI shared control method that can improve the safety and robustness of our system.
2019年03期 v.24 360-370页 [查看摘要][在线阅读][下载 5091K] [下载次数：185 ] |[网刊下载次数：0 ] |[引用频次：34 ] |[阅读次数：1 ] -
<正>Tsinghua Science and Technology (Tsinghua Sci Technol), an academic journal sponsored by Tsinghua University,is published bimonthly. This journal aims at presenting the up-to-date scientific achievements with high creativity and great significance in computer and electronic engineering. Contributions all over the world are welcome.Tsinghua Sci Technol is indexed by SCI, Engineering index (Ei, USA), INSPEC, SA, Cambridge Abstract, and
2019年03期 v.24 371页 [查看摘要][在线阅读][下载 1062K] [下载次数：19 ] |[网刊下载次数：0 ] |[引用频次：0 ] |[阅读次数：1 ] 下载本期数据