Current Issue

• 2017, No. 09　Published:25 September 2017 Last issue    Next issue
  Columns in this issue:

• A dynamic replication placement

  mechanism in cloud storage

  WANG Yan,WANG Jin-kuan

  2017, 39(09): 1581-1587. doi:

  Abstract ( 140 )   PDF (704KB) ( 298 )   　　

  Data replica management is an important part of cloud computing system. When processing massive data in a cloud computing system, the existing algorithms for data storage and resource scheduling do not consider the dynamics and reliability of data replicas．We therefore propose a dynamic replica placement mechanism based on the domain structure. It takes into full account the number and position of data replicas, as well as the  cost of system resources such as memory and bandwidth when a replica is produced. Firstly, according to the information of the replica, the data with high access frequency and long average response time, are replicated, and how to calculate the number of replicas is explained. Secondly, in order to reduce the selection range of nodes of the replica distribution, we propose a dynamic replica placement algorithm, which can choose the range of placement for the replicas according to domain division. Experimental results show that the proposed algorithm can significantly reduce the waste of storage space for the replica with low access frequency, as well as the transmission delay across nodes for the replicas with high access frequency. Besides, it effectively improves the access efficiency to data files in cloud storage systems, the load balance, and the reliability and availability of cloud storage systems.

  A performance management system for

  large-scale system area networks

  CHEN Lin，NAN Yang

  2017, 39(09): 1588-1593. doi:

  Abstract ( 96 )   PDF (663KB) ( 222 )   　　 Review attachment

  For large-scale system area networks, how to effectively monitor the heavy traffic, detect the performance bottlenecks and locate the potential failure points in order to support the optimization of the network performance becomes a new research topic. We firstly put forward a performance management architecture for the system area network called SNPMA, which adopts the loosely coupled hierarchical structure and realizes the automation and operability of the performance management through the collaboration among different layers. Based on the SNPMA, we also propose a network performance evaluation model called NPEM to resolve the problems like the incapability of properly evaluating the performance conditions for network devices and ineffectively predicting the operational status for the whole network in terms of the large-scale system area network. And then we propose a network performance monitoring method called STM based on the adaptive concurrency strategy, which can dynamically adjust the strategies of collecting performance parameters. In the network environment of Tianhe-2, through monitoring the whole network using the STM, the performance indicators of network devices are effectively evaluated, which verifies the validity of the network performance evaluation model NPEM.

  A parallel method for ETL process

  based on agent and activity priority

  CHEN Gang，DU Xin-lin,ZENG Si-feng,AN Bao-ran

  2017, 39(09): 1594-1601. doi:

  Abstract ( 122 )   PDF (679KB) ( 235 )   　　 Review attachment

  ETL is the essential step to obtain high-quality data for data warehouse, and plays an important role in the construction and implementation of data warehouse. Aiming at the deficiency of traditional serial ETL process, we propose a parallel method for ETL based on agent and activity priority. This method first calculates the priority of each ETL activity and then utilizes the agent theory and multi-thread computing techniques to achieve parallel execution of independent ETL activities with the same priority. Experimental results show that this method achieves high speedup when the data volume is large and improves the efficiency of ETL process.

   

  A parallel compressible fluid algorithm based on

  two-dimensional structural grid

  HUANGFU Yong-shuo,LIU Jie，GONG Chun-ye

  2017, 39(09): 1602-1609. doi:

  Abstract ( 102 )   PDF (551KB) ( 215 )   　　 Review attachment

  Based on the two-dimensional/axisymmetric high-precision compressible multiphase flow computational fluid dynamics method  multi-scale simulation code  (MuSiC)-consistent, conservative, all-speed, sharp-interface method (CCASSIM), we propose a parallel area decomposition  method to deal with structural grids. According to different communication ways of boundary data on each processor, blocking and non-blocking communication algorithms are designed respectively. We also develop a MPI / OpenMP hybrid parallel optimization algorithm to reduce communication overhead. Some  numerical experiments are performed on Tianhe-2 supercomputer system with the maximum of grid cells of 1.2 × 1010 and 8 192 processors. The results show that the average parallel efficiency of the program that uses the MPI / OpenMP hybrid parallel algorithm, the pure MPI parallel algorithm with non-blocking communication, and the pure MPI parallel algorithm with blocking communication is 86%, 83% and 77% respectively. Futhermore, good scalability is also observed from the test results of the above three algorithms.

   

  A parallel alignment method for biological

   sequences based on ant colony algorithm

  LI Juan1，TANG De-you1，FU Juan2，3

  2017, 39(09): 1610-1616. doi:

  Abstract ( 131 )   PDF (737KB) ( 257 )   　　 Review attachment

   

  Biological sequence alignment is an important issue in the field of bioinformatics, and the rationality and correctness of alignment results are crucial to the researches based on sequence alignment. It is of great significance to exploit the computational potential with the help of parallel computing to improve alignment efficiency under the premise of ensuring alignment correctness. We propose a parallel alignment scheme based on the ant colony algorithm for the global sequences alignment problem. Aiming at the two most time-consuming steps in the ant colony algorithm, the search of comparison path and the pheromone update, we present a parallel method based on the shared memory model. Experiments on Tianhe II by the OpenMP show that with eight threads in parallel, the speedup can achieve 503, and the longer the sequence is, the better the performance is.

   

   

  A fault-tolerant allocation algorithm for virtual

  machine with controllable redundancy

  ZHANG Yao-guo1,WU Ji-gang2,GAO Ren-fei1

  2017, 39(09): 1617-1626. doi:

  Abstract ( 106 )   PDF (995KB) ( 182 )   　　 Review attachment

  In modern data centers based on virtualization, virtual machine (VM) allocation is the primary consideration for efficient resources scheduling in the cloud. It has been proved that the VM allocation (VMA) problem to minimize the maximum access latency is of NP-hard. But few works on fault tolerant algorithms for the VMA problem are reported, which is very important for the security and reliability of the data center network. We present a heuristic algorithm focusing on minimizing the maximum data access latency. The proposed fault tolerant algorithm constructs the group with controllable redundancy in the set of available VMs, and then selects the VMs from the group and assigns them to the data nodes in the data center network. Extensive experiments on the Tree, VL2, Fat-Tree and BCube, the four commonly used network structures, show that the proposed algorithm can provide between 0 and 200% redundancy, the assignment time of virtual machines for the data nodes is reduced by 67.1% on average when redundancy is between 0 and 40%, and the running time of the algorithm is generally reduced by 12.8%.

   

  A forwarding strategy based on historical

  access records in named data networking

  SHI Ying，ZHANG Xiao-ping,YANG Lei

  2017, 39(09): 1627-1637. doi:

  Abstract ( 120 )   PDF (1443KB) ( 200 )   　　 Review attachment

  As the instance system of information center networking (ICN), the named data networking (NDN) attracts more and more attention in academia. NDN changes the network service from “destination to destination” to “retrieving data of given names”. NDN is more convenient for consumers to share data and it gets a substantial increase in the rate of repeated utilization of data. We focus on the forwarding of NDN routing nodes and propose a forwarding strategy based on historical access records called HRF, which can find out whether there is target data stored by the neighbor nodes in the local network, and then access the target data in the local network efficiently and rapidly. Through a series of typical scenarios, we validate the feasibility and effectiveness of the HRF.

   

  A load redistribution strategy based on

  dynamic information in cascading process

   

  LIANG Zhuo-qian1,2,FU Dan-long2，DENG Yuan1

  2017, 39(09): 1638-1644. doi:

  Abstract ( 115 )   PDF (955KB) ( 271 )   　　 Review attachment

  Based on cascading failures of the network, we investigate the same process of interdependent heterogeneous networks such as ER-BA, and propose a load redistribution strategy based on real-time processing ability of nodes. Simulation tests show that the overall robustness of the network is mainly affected by the BA network. The connection strength should be reduced when the cascading process is initiated. The load  redistribution strategy based on the betweenness of nodes can easily lead to the failure of the central nodes and reduce the overall robustness of the network. The proposed strategy can optimize the overall robustness of interdependent heterogeneous networks.

   

  A WSNs Bayesian localization and tracking algorithm

   based on Kullback-Leibler divergence filtering

  MENG Fan-kun,JU Yong-feng,WEN Chang-bao

  2017, 39(09): 1645-1652. doi:

  Abstract ( 116 )   PDF (994KB) ( 247 )   　　 Review attachment

  In order to achieve high tracking precision of moving target positioning in wireless sensor networks (WSNs) with low energy consumption and bandwidth consumption, we propose a WSNs Bayesian localization and tracking algorithm based on Kullback-Leibler divergence filtering. Firstly, we use the Gaussian and Wishart distribution without considering the speed limits and restrictions of movement direction to construct a mobile localization Bayesian state evolution model for WSNs, as well as a moving target positioning observation model based on the path loss model. Secondly, we use the Kullback Leibler to construct a divergence filtering error calculation model, which can estimate the position of the goal of the mobile node through activating the surrounding nodes. The recursive probability calculation process we designed integrates prediction and updates process, and realizes synchronous target localization and tracking. Simulation results show that the proposed model has advantages in tracking accuracy and resource saving.

  A secure steganography algorithm

  based on wavelet contrast and modulus

  WEI Wei-yi,ZHU Qiang-jun

  2017, 39(09): 1653-1656. doi:

  Abstract ( 101 )   PDF (632KB) ( 216 )   　　 Review attachment

  To obtain a larger embedding capacity and better imperceptible stego-images, we propose a new image steganography algorithm using wavelet contrast and modulus operation based on the visual characteristic that humans are not sensitive to rapid changing areas and dark areas. Firstly, the carrier image is divided into blocks with a fixed size and the wavelet contrast is calculated. Secondly, the embedding depth of each block is determined according to the wavelet contrast. Finally, secrete data is embedded into the image through modulus operation. Experimental results show that the proposed approach provides both larger embedding capacity and higher image quality.

   

  Software interactive performance anomaly analysis

  HE Jia-quan1,CHEN Yu2

  2017, 39(09): 1657-1664. doi:

  Abstract ( 108 )   PDF (645KB) ( 208 )   　　 Review attachment

  The development of personal computers and mobile devices, as well as modern operating systems, has promoted application software developing to a new level. Multithreading program development is then introduced to optimize software user experience. While it makes full use of hardware and shortens software response latency, it complicates software development and makes it more difficult to analyze and debug. We look into the existing work regarding interactive performance anomaly analysis, broaden the range of previous known interactive performance anomalies, and improve the existing analyzing model to meet the performance anomaly analysis needs of application software in Linux operating system. Based on the mentioned work, we design an interactive performance anomaly analyzing system on Linux, and using popular software such as Google Chrome and GNOME Nautilus as the objectives we analyze some interactive performance anomalies in practice.

  A data model of sharing system based on Fibrations theory

  MIAO De-cheng1,XI Jian-qing2,DAI Jing-guo1

  2017, 39(09): 1665-1674. doi:

  Abstract ( 93 )   PDF (678KB) ( 203 )   　　 Review attachment

  There are some drawbacks for traditional modeling methods of data model of sharing system in analyzing semantic properties and describing semantic behaviors. Aiming at the problems mentioned above we present a data model of sharing system based on Fibrations theory. We contribute in the following two aspects. Firstly, we accurately analyze semantic properties by combining algebras methods with truth functor, lifting preserving-truth and comprehension functor, and formally depict semantic behaviors by combining co-algebras methods with equation functor, lifting preserving-equation and quotient functor. Secondly, in the framework of Fibrations theory we construct parameterized recursive and co-recursive operations on complex inductive and co-inductive data structure to abstractly describe inductive and co-inductive rules with universality, and briefly introduce applications of Fibrations theory by examples. Compared with traditional methods such as category theory, the Fibrations theory of brief descriptions and flexible expansibility can accurately analyze semantics properties, formally describe semantic behaviors of data model of sharing system, and abstractly depict inductive and co-inductive rules with universality of complex data structures.

  A satellite image denoising method based

  on MPSO and dictionary learning

  WANG Xiao-yan1,2，CHI Tian-he1

  2017, 39(09): 1675-1681. doi:

  Abstract ( 98 )   PDF (851KB) ( 190 )   　　 Review attachment

  Online dictionary learning (ODL) updates all dictionary atoms and it is difficult to estimate the optimization direction, so the accuracy is decreased. Aiming at this drawback, we propose a method of online dictionary learning based on modified particle swarm optimization (MPSO). On the basis of ODL, the algorithm optimizes the gradient descent function in the iterative process of dictionary learning.  A special atom is selected by a rule in the dictionary-updating stage, which is linearly represented by the other atoms in the dictionary. The coefficient of the linear representation is the position of the particles in the MPSO, which is introduced to eliminate the particles by their fitness while leaving the more suitable particles in the next iteration. Furthermore, intermediate variables, which carry the prior reference data, are introduced into the MPSO to guide the optimization direction, so that the  direction is constrained to improve the accuracy and effectiveness of the dictionary. Experimental results confirm that compared with the other three algorithms, this algorithm has a better performance in noise suppression and improves the performance of large-scale data computation.

   

  Independent component analysis-based channel selection

  for high-accuracy classification of N200 and P300

  LI Wen-xuan1，LI Wei2，LI Meng-fan1，LIU Cheng-yong1

  2017, 39(09): 1682-1690. doi:

  Abstract ( 133 )   PDF (903KB) ( 485 )   　　 Review attachment

  Since EEG signals have individual difference and are vulnerable to noise and artifacts, we propose an independent component analysis (ICA)-based method for the selection of optimal feature channels. This method applies the ICA to decompose channels’ data to N200, P300, ocular artifacts and other physiological signals. Whether a channel is suitable for feature extraction is decided by the influence of those signals that mentioned above to this channel. We apply our method and three other commonly used methods for feature channel selection to twelve subjects’ brain signals, and recognize N200 and P300 potentials. We find that our method achieves a 93.10% accuracy on average and it is 7.27%, 1.07% and 75.96% higher than the average accuracy of the other three methods respectively. We fit a relation curve between ICA weight and channel selection threshold based on the least square method, and obtain a high classification accuracy when predicting the optimal channels and recognizing the potentials from another three new subjects' data, which means that this prediction method has universality.

  A lace retrieval method based on hierarchical

  matching and multiple features

  CAO Xia1,LI Yue-yang1,LUO Hai-chi2,JIANG Gao-ming1,CONG Hong-lian1

  2017, 39(09): 1691-1699. doi:

  Abstract ( 115 )   PDF (1101KB) ( 190 )   　　 Review attachment

  Since the efficiency of the lace retrieval method based on image texture features is low, and in order to extract the effective texture features for lace identification, we propose a lace retrieval algorithm containing multiple features fusion through hierarchical matching. Firstly, we process the texture image by the Canny operator and obtain the Canny color image and the gray level-gradient co-occurrence matrix (GGCM). Secondly, energy, average gradient, average grayscale, correlation and other statistical characteristics are used for texture description. Finally, the extracted texture features are matched with geometry features and speeded up robust features (SURF) hierarchically, so the fusion of multiple features under hierarchical matching is realized to compensate for the deficiency of any single matching method and verify the correct rate of the retieval method used in the lace library. Experimental results indicate that the performance of the proposed method is better than other methods, which has a stronger ability of texture identification, and can achieve lace retrieval effectively and improve the reliability and accuracy of image retrieval.

  Image registration based on

   multi-scale corner with SIFT descriptor

   

  SU Pei-feng1,HUANG Shi-qi2,WANG Yi-ting1,LIU Dai-zhi1

  2017, 39(09): 1700-1705. doi:

  Abstract ( 99 )   PDF (1067KB) ( 200 )   　　 Review attachment

  The corners of images are rich in image structure information and widely used in image registration. The Harris algorithm is a classical corner extraction algorithm, and Harris corners are invariant to image rotation but sensitive to image scale change, which limits its application in image regis-
  tration when the image scale changes. We propose a multi-corner extraction method based on the feature extraction process of the scale invariant feature transform (SIFT) algorithm, which is insensitive to image rotation and scale changes. We use the SIFT descriptor to describe the multi-scale corners to realize image registration. Experiments on optical and SAR images show that compared with the SIFT and Harris algorithm, the proposal reduces the registration time by more than 40% and improves the utilization of the characteristic points by more than one time while maintaining the registration precision.

  
   

  An object recognition method combining

  saliency detection and bag of words model

  LI Wei-sheng,CHEN Xi

  2017, 39(09): 1706-1713. doi:

  Abstract ( 91 )   PDF (931KB) ( 191 )   　　 Review attachment

  Given that the bag of words model is quite sensitive to background noise and that visual words in the background are not relevant to objects, we propose an object recognition method which combines saliency detection with the bag of words model. Firstly, the region of interest from the original image is adaptively gained by using the graph-based visual saliency (GBVS) algorithm and the AC algorithm. The combination of the two detection methods can avoid incomplete region of interest. Secondly, we extract local features from the region of interest by using the scale invariant feature transform (SIFT) descriptor. Then, we use the peak density clustering algorithm to classify the features and generate a visual dictionary histogram by clustering local features. Finally, we employ the support vector machine (SVM) classifier to classify and recognize objects. Experiments on PASCAL 2007 and MSRC-21 databases verify the effectiveness of this method. Experimental results show that the proposed method can effectively improve the performance of object recognition.

  A new  human model generation algorithm

  based on orientation deformation on primitives

  HUANG Xin,MA Xi-rong,ZHAO Zi-ping

  2017, 39(09): 1714-1720. doi:

  Abstract ( 84 )   PDF (836KB) ( 222 )   　　

  We present a human model generation algorithm based on orientation deformation on primitives. By using the primitive curve spots and orientation deformation, we can obtain human primitives of different human parts. At the same time, it can also eliminate displacement when some simulated functions are used. Meanwhile, different human models are generated through primitive mesh generation and smooth connection. This method can not only improve the reusability of generated human primitives, but generate some specific models via the combination of different types of primitives. Experimental results show that compared with traditional methods, the models generated by the proposed method are more real and have less difference with the template model. Furthermore, the generation efficiency of human models can be further improved because of the reusability of primitive models.

  An improved video zero-watermarking

  algorithm in pseudo 3D-DCT domain

  JIANG Ye-qian,SONG Chun-lin

  2017, 39(09): 1721-1728. doi:

  Abstract ( 75 )   PDF (996KB) ( 226 )   　　 Review attachment

  In order to improve the robustness of the video watermarking algorithm, we propose an improved video zero-watermarking algorithm in pseudo 3-D discrete cosine transform (3D-DCT) domain. In this scheme, we first select key frames through the inter-frame Euclidean distance algorithm, and then apply the three frames difference algorithm to extract the moving object of the key frame. Afterwards the moving object of the key frame is segmented as blocks, and the pseudo 3D-DCT is performed on the blocks. After the sequences of video eigenvalues are constructed by AC coefficients, we embed watermark signals into the sequences of eigenvalues to structure zero-watermark. Finally, the zero-watermark is registered in a database. Experimental results show that the proposed algorithm is invisible and robust against most attacks, including noise filtering attacks, frame cropping, rotation, scaling and so on.

  Consensus analysis for heterogeneous multi-agent

  systems with immeasurable velocity

  ZHU Mei-ling，ZHAO Rui，XU Yong

  2017, 39(09): 1729-1735. doi:

  Abstract ( 110 )   PDF (695KB) ( 210 )   　　 Review attachment

  With the wide application of heterogeneous systems in actual systems in recent years, consensus research for heterogeneous multi-agent systems becomes a hot research area. For heterogeneous multi-agent systems composed of first-order agents with bounded control input and second-order agents with immeasurable velocity, we propose a leaderless control protocol and a leader-follower control protocol for heterogeneous multi-agent systems under undirected network topology. Some sufficient conditions for consensus problem are obtained by applying the graph theory and LaSalle invariance principle. Finally, simulations verify  its validity.