Author Login Editor-in-Chief Peer Review Editor Work Office Work

05 January 2008, Volume 34 Issue 1
    

  • Select all
    |
    Degree Paper
  • LI Feng; WEI Jun-yin
    Computer Engineering. 2008, 34(1): 1-3. https://doi.org/10.3969/j.issn.1000-3428.2008.01.001
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    On the base of auditory neural system, the on-center off-surround network model on processing of the sound wave is presented. In the network model, the neuron in the input layer excites the corresponding neuron in the output layer meanwhile it inhibits the neighborhood neuron. In the output layer, the neuron production auto-attenuates at a definite speed too. The mathematic equation and the learning algorithm on the neural network are also presented. The learning algorithm convergence is proved. The network has its advantage in sound wave processing. In addition, it can filter the noise, search the significance frequency segments and transform the speech physical spectrum to the biophysics action model. The “channel suppresser”, the human acoustical system special phenomenon, is explained based on the model.
  • QI Li; LIU Yu-shu
    Computer Engineering. 2008, 34(1): 4-6. https://doi.org/10.3969/j.issn.1000-3428.2008.01.002
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    To accelerate the training of Support Vector Machine(SVM), this paper proposes a new fuzzy support vector machine based on two stage clustering. After applying density clustering algorithm to the training samples of each class, the initial reduced training set is generated based on those clustering edge-samples. According to its distance towards separating plane, the neighbor radius of each sample is calculated. The new reduced training set is generated based on adaptive density clustering algorithm. Experimental results show that Fuzzy Support Vector Machine(FSVM) based on two stage clustering can get better classification results while reducing the training time greatly.
  • XU Gang; DING Quan-long
    Computer Engineering. 2008, 34(1): 7-9. https://doi.org/10.3969/j.issn.1000-3428.2008.01.003
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper proposes a Prediction-based Longest Queue First (PLQF) scheduling algorithm. The scheduler considers not only the queue length, but also the incoming traffic. Based on that information, the resource is allocated to the user with the highest probability of overflow. The scheduler can adjust the queue lengths in advance to cater for the incoming traffic, targeting to minimize CLR and improve the buffer utilization. Theoretical analysis shows that the PLQF algorithm can lead to a lower CLR than the conventional LQF algorithm. Simulation results show the PLQF algorithm can reduce CLR by 10%~60%.
  • WANG Hui-fang; ZHU Zhi-qiang; SUN Lei
    Computer Engineering. 2008, 34(1): 10-13,1. https://doi.org/10.3969/j.issn.1000-3428.2008.01.004
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    There is no reliable trust system in Internet nowdays. It attracts more and more attentions how to express and manage trust relations in a distributed and disordered network system. This paper analyzes the define of trust and logic model of trust firstly, furthermore researches especially on trust evaluation model, trust model, and trust management of distributed network systems. Also it discusses the problem of trust management in distributed network systems.
  • WANG Hong-mei; ZENG Yuan; ZHAO Zheng
    Computer Engineering. 2008, 34(1): 14-16. https://doi.org/10.3969/j.issn.1000-3428.2008.01.005
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Privacy regulations may prevent parties from sharing their data. A privacy-preserving EM learning on horizontally partitioned data with missing data (PPHI-EM) is proposed to make parties share their data under privacy. Each party owning confidential data disassembles the likelihood function, replacing the expected value of all the statistics by the nonexistent value. Based on AMS-EM, the network structure is improved to convergence by iteration cycle. The intersection and union of directed edges of structure are computed with security directed edge statistic algorithm, in which pohlig-hellman encryption algorithm is used. The intersection is regarded as the initialization of structure. The edge in union other than that in intersection is put into the structure one by one. The proposed method alternates between the iterations that optimizes the parameters for the current model candidate and the iterations that searches for a different model. Parameters are solved with security matrix sum algorithm by enactment of proper weight sum. It lies on the values of scoring function that the edge is remained or not in this aggregation. Experimental results show its effectiveness.
  • KANG Ye-wei; HUANG Ya-lou; SUN Feng-chi; YUAN Jing
    Computer Engineering. 2008, 34(1): 17-19,2. https://doi.org/10.3969/j.issn.1000-3428.2008.01.006
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Simultaneous Localization And Mapping(SLAM) is a necessary prerequisite to make robot autonomous, which is a harder research topic than localizing or mapping. A Rao-Blackwellised Unscented Kalman Filter(RBUKF) based SLAM method is presented which uses the Scaled Unscented Transformation(SUT) to sample the Sigma points for robot operating in plain static environment. With the same computing complexity, RBUKF can avoid linearization error introduced in the Extended Kalman Filter(EKF) filter, which can induce the final map error. The experimental result of the method based on the public dataset is better than the EKF based method according to the precise of the final estimated map.
  • DING Shi-fei; SHI Zhong-zhi; JIN Feng-xiang
    Computer Engineering. 2008, 34(1): 20-22. https://doi.org/10.3969/j.issn.1000-3428.2008.01.007
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    On the basis of analysis and studies on modeling method of Partial Least squares (PLS), a novel algorithm of information pattern recognition based on Nonlinear Iterative Partial Least Squares (NIPLS) is set up in this paper. This algorithm reasonably combines feature extraction with classifier design, and has more advantages than classical algorithm of Fisher Discriminant Analysis (FDA) and Bayes Discriminant Analysis (BDA), such as simplicity and robustness, clearly qualitative explanation, and applies it to classification recognition of land quality. The simulation results show that the algorithm has better recognition effect than FDA and BDA. It is powerful for multicollinearity, particularly when the number of predictor variables is large and the sample size is small, and provides a novel and efficient algorithm for pattern recognition.
  • ZHANG Hong; HE Ye-ping; SHI Zhi-guo;
    Computer Engineering. 2008, 34(1): 23-25. https://doi.org/10.3969/j.issn.1000-3428.2008.01.008
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper presents an improved data refinement proof rules. These rules are formulated on relations, including a global state to describe all possible input and output for programs, allowing non-trivial initialization, allowing both forward and backward simulations. These rules are also applicable when the non-determinism resolution in the concrete model is later than in the abstract model. In addition, it uses a simple example to illustrate the application of the rules in Isabelle theorem prover.
  • TONG Wei-min; LI Yi-jun; SHAN Yong-zheng
    Computer Engineering. 2008, 34(1): 26-29. https://doi.org/10.3969/j.issn.1000-3428.2008.01.009
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper presents wavelet method and ARMA model for time series data mining. According to the wavelet denoising and wavelet decomposition, the hidden period and the nonstationarity existed in financial time series are extracted and separated by wavelet transformation. The characteristic of wavelet decomposition series is applied to BP networks and an Autoregressive Moving Average(ARMA) model. Finally, wavelet reconstruction is used to realize time series forcaseting. It shows that the proposed method can provide more accurate results.
  • YANG Wan-kou; REN Ming-wu; YANG Jing-yu
    Computer Engineering. 2008, 34(1): 30-33. https://doi.org/10.3969/j.issn.1000-3428.2008.01.010
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Chain code is very useful to object description in digital image, so area algorithms based on chain code are used widely. The paper analyzes and compares the chain codes based area algorithms from the algorithm principles, speed and precision, and gives the best algorithm for the small object.
  • YIN Pan; XU Qiang
    Computer Engineering. 2008, 34(1): 34-36,3. https://doi.org/10.3969/j.issn.1000-3428.2008.01.011
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    In the field of urban planning, it is important for remote sensing data to use vector. By analyzing the characteristics of raster and vector data, complex mathematical morphology is improved. Through the wavelet analysis, boundary processing and amending work, a better vector map can be gained with different types of land using, which improves the veracity rationality and science of urban planning.
  • ZHENG Pei-chao; SONG Han-tao; XU Xiao-mei
    Computer Engineering. 2008, 34(1): 37-39. https://doi.org/10.3969/j.issn.1000-3428.2008.01.012
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Water-filling distribution algorithm is the optimal algorithm for OFDM systems to maximize the overall bit rate under overall power and bit error rate constraints. Dichotomy combined Lagrange-multiplier method is used to estimate the target value. The algorithm has slow convergence when the estimated value is far from the target. This paper presents a mixed dichotomy to solve the problem. Simulation results show that the proposed algorithm has faster convergence than dichotomy.
  • XU Qian-fang; XIAO Bo; GUO Jun
    Computer Engineering. 2008, 34(1): 40-42,4. https://doi.org/10.3969/j.issn.1000-3428.2008.01.013
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Currently those algorithms to mine the alarm association rules are just able to obtain the association rules among the frequently occurring alarm events, limiting to the minimal support. To address this problem, a new algorithm based on the statistical correlation is proposed to discover the association rules from both high-frequency and low-frequency alarm events with the high correlativity and the high confidence. And an incremental algorithm is proposed to discover new rules in an updated database. Experimental results have demonstrated that the algorithms are efficient and accurate.
  • LI Hui; WANG Hang-yu; WANG Hou-xiang
    Computer Engineering. 2008, 34(1): 43-46. https://doi.org/10.3969/j.issn.1000-3428.2008.01.014
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Aiming at the multi-domain and dynamic characteristics of the access control in gird, the shortage of traditional access control strategy is demonstrated. The RBAC model is functionally branched out and the G_ERBAC model is presented. In G_ERBAC model, the environment entity and the concept of domain management are introduced. With its user-role assignments performed according to dynamic role reflection rules between domains, and role-permission adjusted by environment-based dynamic security engineering machine, the G_ERBAC model provides user-active role-active permission associated dynamically, and it is more suitable for grid.
  • SHI Xiao-hua; JIN Mao-zhong
    Computer Engineering. 2008, 34(1): 47-49. https://doi.org/10.3969/j.issn.1000-3428.2008.01.015
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper introduces an optimization technology of code generation, namely lazy code selector, in just-in-time compiler, and demonstrates the details of designation and implementation of this technology in Intel’s high-performance J2ME JVM, XORP. Other lightweight optimizations in XORP have also been introduced. EEMBC is used to evaluate the performance of XORP with other JVMs, and the impact of different optimizations in XORP. With the lazy code selector, XORP is more than fifteen times faster than KVM, and faster than other commercial high-performance JVM.
  • HE Jing; WANG Xin-sai; LI Qing-hua
    Computer Engineering. 2008, 34(1): 50-52. https://doi.org/10.3969/j.issn.1000-3428.2008.01.016
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper, based on imaging model, uses illumination model and atmosphere model to research the relation among date, geographical position and sun position, and the relation between sunlight and sun position under condition of the different climate. Then according to the climate, geographical position and date, sunlight of the scene surface can be determined. By already-known DEM, viewpoint and reference map, terrain optics image under condition of the different climate and time can be realistically simulated.
  • LI Xiao-mao; WANG Zhi-feng; TANG Yan-dong
    Computer Engineering. 2008, 34(1): 53-55,5. https://doi.org/10.3969/j.issn.1000-3428.2008.01.017
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper integrates the detected object shape represented by shape restraint equation and general active contour model into an unified energy functional generating the shape preserving active contour model. In this model, a specific class contour shape is represented as the zero level line of some certain parametric level set function. This model can not only detect the given object correctly but also characterize the object shape quantificationally via these parameters. In addition, it builds an energy functional for long straight bar detection using the proposed model. It deduces the corresponding Euler-Lagrange ODE) of its shape parameters and implements them using level sets method. It also deduces that a special case of the long straight bar detection model can be used for the linear skyline (sea level) detection. Numerical experiments present this model not only can detect the long straight bar correctly but also is robust to noise, deformity and occlusion.
  • Software Technology and Database
  • CAO Zhong-sheng; XUE Mei-yan
    Computer Engineering. 2008, 34(1): 56-58. https://doi.org/10.3969/j.issn.1000-3428.2008.01.018
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper proposes an approach to software classification in fine-grain based on SLIQ algorithm, uses testing tools and writes API’s to get both inner and outer attributes of software, cleans these data and abstract classification rules of software, and builds a classifying model applied in DBMS classification.
  • WANG Cheng-yi; ZHAO Zhong-ming
    Computer Engineering. 2008, 34(1): 59-60,6. https://doi.org/10.3969/j.issn.1000-3428.2008.01.019
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    LIDAR data are new data source, and they generate a high spatial resolution “point cloud”. To use it, the first step generally is to translate it to Digital Surface Model (DSM). The traditional methods are limited in terms of generation of DSM in urban areas because the urban environment is extremely complex with many man-made objects of different features and heights, and those methods do not consider conditions when data are missing. DSM generated using these traditional interpolation methods cannot achieve the purpose needed for urban applications, such as precision. This paper presents a new method to generate DSM considering characteristics of cities and traits of LIDAR sensors. The DSM generated using this method has better quality than that using transitional ones.
  • ZHAO Su; LI Xiu; LIU Wen-huang
    Computer Engineering. 2008, 34(1): 61-63. https://doi.org/10.3969/j.issn.1000-3428.2008.01.020
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper presents an evaluation model for the text classifier. The reliability of classifying result of a classifier is computed according to its learning result and naive Bayesian. Based on the performance evaluation model, Performance Bagging(PBagging), an improved text classification algorithm is proposed. In the algorithm, the reliability is served as the weight of subclassifier’s result when using Bagging, an ensemble learning method. Using SVM as the sub model, it applies the PBagging algorithm to classify news corpus in kyodo news agent, the result shows that PBagging performs better than Bagging with more accuracy.
  • ZHANG Zong-ping
    Computer Engineering. 2008, 34(1): 64-65,6. https://doi.org/10.3969/j.issn.1000-3428.2008.01.021
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    In data mining, IUA algorithm has a problem that the frequent itemsets are not minined completely. Apriori algorithm in the analysis, such as classical correlation rules IUA algorithms excavation algorithms based on the results presentes an update on recent excavations algorithm called HIUA. HIUA algorithms absorbed Apriori algorithm and the advantages IUA algorithms, the smallest change in the support for and based on the results of recent excavations conditions from the pending option sets generated minimal consideration given complete sets of innovative procedures, thereby enhancing the efficiency of algorithms.
  • LIU Peng; NIU Qiang; CHEN Dai; ZHANG Bao-hui
    Computer Engineering. 2008, 34(1): 66-68. https://doi.org/10.3969/j.issn.1000-3428.2008.01.022
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper analyzes why priority-reversal problem occurs in μC/OS-II, and the limitation of μC/OS-II. An improved priority ceiling protocol(IPCP) is introduced, which including its principle, basic content. The protocol solves priority reversal problem, has other good characteristics, such as preventing system dead lock, etc. The approach of implementation of IPCP by modification of kernel of μC/OS-II is proposed.
  • LIU Bo; LEI Gang-yue; YANG Lu-ming; DENG Yun-long
    Computer Engineering. 2008, 34(1): 69-71. https://doi.org/10.3969/j.issn.1000-3428.2008.01.023
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Integrity constraint is important to make data certain in relation database, but there is a larger amount of uncertain and inconsistent information that is valuable and useable. Combined with probabilistic database theory, this paper gives a new query plan aiming at inconsistent database. It uses the constraint methods including union, product, subtraction, selection, projection and link to repair inconsistent data effectively. Its probabilistic calculation with four elements and probabilistic query rewriting can overcome shortcomings of inconsistent databases and decrease conflict of data.
  • ZHANG Lu-di; ZHU Qun-xiong
    Computer Engineering. 2008, 34(1): 72-74,7. https://doi.org/10.3969/j.issn.1000-3428.2008.01.024
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper presents an approach of interaction modeling in virtual reality. A new modeling tool VRTool is designed, and topology model(VRTP) and Visual Training Navigation model(VTN) are proposed to solve the problems of data source of location coordinates and linear pace mutation, which can set up interaction relation of man-machine and three-dimension navigation movement conveniently. The experimental result demonstrates the usability of this approach, and a new thought for interaction modeling is provided to the development of VR modeling tool.
  • ZHANG Jia-kou; ZENG Guo-sun;
    Computer Engineering. 2008, 34(1): 75-77. https://doi.org/10.3969/j.issn.1000-3428.2008.01.025
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    After benefiting from the biodiversity characteristic and biology taxonomic classification, this paper proposes a classification scheme for the Distributed Computing Software Resources (DCSR) referring to the characteristic of the distributed computing software resources. DCSR can be clearly described and classified, also is matched to personal demand optimally. The result shows that the technology can identify 70%~80% ftp software, and preliminary solution is given for distinguish heterogeneous resources problem.
  • ZHANG Jie; YUAN Hong-liang; GUO Chang-guo; ZHANG Zhu-xi ; GONG Hai-lin
    Computer Engineering. 2008, 34(1): 78-79,8. https://doi.org/10.3969/j.issn.1000-3428.2008.01.026
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    An imperfect event forwarding algorithm based on Bloom Filter technology, namely FBBF algorithm, is proposed. FBBF algorithm increases the overhead of event message size and network communication in order to improve the speed of event forwarding in intermediate broker. The time cost of FBBF algorithm is independent of the scale of subscriptions, so FBBF algorithm has good scalability and can be efficiently applied to large scale publish/subscribe system.
  • DENG Jing-jing; YE Xiao-jun
    Computer Engineering. 2008, 34(1): 80-82. https://doi.org/10.3969/j.issn.1000-3428.2008.01.027
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    K-anonymization is an important approach to protect data privacy in data publishing scenario. Like K-D tree for multidimensional K- anonymity, this paper proposes, an implementation of R tree in which each record is considered as a point in d-dimensional space of the attribute. Instead of dividing the region into pieces, the nearby rectangles are grouped into parent minimal bounding rectangles and forms disk blocks. Experiment results by modifying several parameters show that the algorithm can handle higher dimensionality compared with grid file or k-d tree.
  • YIN Hong-zhang; CAO Xian-bin;
    Computer Engineering. 2008, 34(1): 83-85. https://doi.org/10.3969/j.issn.1000-3428.2008.01.028
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper proposes a new method of mining social network by combining content comparability and sequencing contained in the chat data. The method infers primary social network using some heuristic rules, then reinforces the network by users clustering. Experimental data show that the method has a good performance.
  • HU Chang-jun; DING Liang; CHANG Xiao-dong; LI Jian-jiang
    Computer Engineering. 2008, 34(1): 86-89,9. https://doi.org/10.3969/j.issn.1000-3428.2008.01.029
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper proposes a parallel modeling method to improve design decision and reduce time-consuming software development. It describes how to customize a series of models for shared and distributed memory application by using UML extension mechanisms, and how to support real applications by combing these basic models. The method maintains the features of UML such as facilitating understanding, customization, guideline for practical application and extension from sequential modeling to parallel modeling. The usefulness of this method is demonstrated by modeling a real parallel application.
  • HOU Jun; WANG Yong-jian; QIAN De-pei; BAI Yue-bin; WANG Ke
    Computer Engineering. 2008, 34(1): 90-93. https://doi.org/10.3969/j.issn.1000-3428.2008.01.030
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper proposes a grid-based federated digital library solution. In the solution, digital library data resource node is created by encapsulating the digital library asset with services. A digital library asset federation is built to integrate depot nodes with grid concept to transparently access the data assets of the integrated nodes. The federation can provide a search-based dynamic customization engine for digital library instance to meet the requirements of building professional and personal digital libraries.
  • SUN Xia; YANG Yun
    Computer Engineering. 2008, 34(1): 94-96. https://doi.org/10.3969/j.issn.1000-3428.2008.01.031
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Software-architecture-level maintenance evaluation based on scenario mainly relies on two technologies: elicitation of the scenarios and impact analysis in architecture level. This paper focuses on the impact analysis in architecture level. It studies the challenges that impact analysis faces, and presents the solution. The study in this technology can improve the efficiency of the many software-architecture-level evaluation methods, and can improve system maintenance.
  • LIU Ting-zhang; ZHU Hong; YE Bing; MA Jing
    Computer Engineering. 2008, 34(1): 97-99. https://doi.org/10.3969/j.issn.1000-3428.2008.01.032
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Most of the data front-end processing in supporting software package for Rapid Prototyping(RP) system is based on identified RP technology, and cannot serve as a common platform for all RP technologies because of the difficulties in redevelopment, integration and standardization of the technologies. This paper proposes a real-time method to generate hatching lines. The generation is repositioned between two layers building, and its algorithm is designed for fast processing in the memory. The method is generalized to suit for global use and it can be embedded in the general supporting software package as a common component for all RP technologies.
  • DUAN Jie-feng; LI Qing; ZHOU Mei-lian; CAI Qun
    Computer Engineering. 2008, 34(1): 100-101,. https://doi.org/10.3969/j.issn.1000-3428.2008.01.033
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper discusses the distributed parallel computing for LBGK D2Q9 model which is based on lattice Boltzmann method, and studies the influence of distribution of two-dimensional data and its communication operation on parallel efficiency. Lots of numerical experiments on "Ziqiang3000" supercomputers are done and experimental results show that LBGK model has good parallelism, scalability and parallel efficiency, which matches the theoretical analysis.
  • JIANG Lei; HUANG Guang-jun
    Computer Engineering. 2008, 34(1): 102-104. https://doi.org/10.3969/j.issn.1000-3428.2008.01.034
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    The traditional lossless compression algorithms are aim for higher compress ratio, but they can not meet the requirement of the network environment. This paper raises an adaptive algorithm with adjustable compression ratio by parameter to meet the requests of different enviroments. The experimental result shows that the improved algorithm is more efficient than the tradional one in the process of remote transmission.
  • Networks and Communications
  • ZHANG Yan-mei; YU Zhen-wei; CAO Huai-hu; JIN Xin
    Computer Engineering. 2008, 34(1): 105-107. https://doi.org/10.3969/j.issn.1000-3428.2008.01.035
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper proposes a routing algorithm for composed service. Routing tragedies as LDT and LDT-P algorithms are designed to fit the line and parallel structure composition respectively. Composition agent is responsible for searching components and generating a least delay tree. The processing ability of the nodes and the bandwith of links are considered comprehensively while selecting a path for composed service. Simulation shows that LDT and LDT-P algorithms have the feature of higher composition success rate and shorter path length, compared with the currently routing algorithms.
  • WANG Jian-bai; CHEN Feng; LOU Yi-hua
    Computer Engineering. 2008, 34(1): 108-110,. https://doi.org/10.3969/j.issn.1000-3428.2008.01.036
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper designs and implements a Per-flow handover protocol based on mobile IPv6. It adds two messages(flow binding update, flow binding acknowledgement) and a flow binding list to mobile IPv6, and implements flexible data flow redirection by modifying Linux network kernel modules. The protocol provides more functions for intelligent handover policies of high layers. The experiments demonstrate that the Per-flow handover improves the performance of transmission and uses the network resource more efficiently.
  • SHI Zhen-yu; LI Zhao; XU Ke
    Computer Engineering. 2008, 34(1): 111-113. https://doi.org/10.3969/j.issn.1000-3428.2008.01.037
    Abstract ( ) Download PDF ( )   Knowledge map   Save

    Existing schemes in IEEE 802.11 network can only resolve the problems of inner ESS domain, but cannot effectively resolve that of inter ESS domains because the handoff between two ESS domains involve network layer operations. This paper proposes a smooth handoff scheme based on virtual interface in wireless network. This scheme applies a time-sharing policy on physical wireless interface, which enables mobile station to establish new wireless link with the right new access point and to keep normal communication in present wireless link with the previous access point at the same time, so it can perfectly resolve the problems of packet latency and loss ratio for both inner ESS domain handoff and inter ESS domain handoff. Simulation results on NS2 confirmes that it can satisfy the demand of interactive real-time applications in wireless mobile networ

  • XU Hai; XU Tao
    Computer Engineering. 2008, 34(1): 114-116. https://doi.org/10.3969/j.issn.1000-3428.2008.01.038
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Anycast is a new type of network service, and has been defined as a standard communication model of IPv6. With anycast service, users can reach the appropriate one of the servers that accepts datagram for the anycast address. With the increment of applications need of anycast service, the issue of anycast routing becomes a significant subject to study. This paper gives the research on anycast service and algorithms done before, and introduces an algorithm based on genetic algorithm, which uses double populations. Experimental simulation shows the improved algorithm can effectively speed the up searching process, boost the chance of achieving the best path, and improve the quality of network service.
  • LIN Tong; SHEN Xiang-jun; QIAN Hua-lin
    Computer Engineering. 2008, 34(1): 117-119,. https://doi.org/10.3969/j.issn.1000-3428.2008.01.039
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Multicast islands are always interconnected by tunnels. This method needs to append extra control information to each packet and introduces high overhead to procedures such as head processing and data transmission. In order to archive efficient global multicast, this paper propose a substitute which is based on address and port translation. It enumerates existing tunnel types and describes the detail mechanisms of the MAPT centralized approach, also presents simulation results and related discussion.
  • CHEN Hua-sheng; XU Rui; LI Wei-hua
    Computer Engineering. 2008, 34(1): 120-121,. https://doi.org/10.3969/j.issn.1000-3428.2008.01.040
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Flow label is introduced in IPv6 protocol, and used to provide routers for QoS implementation purpose of forwarding packets. This paper focuses on the demand of IPv6 multimedia multicast communication, and proposes an IPv6 multicast communication scheme of flow label classification providing QoS using a hybrid approach. The scheme with multimedia multicast communication is tested on IPv6 test network, and satisfactory QoS results are gotten.
  • LI Jian; WANG Jian-xin
    Computer Engineering. 2008, 34(1): 122-124. https://doi.org/10.3969/j.issn.1000-3428.2008.01.041
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper proposes a differentiated back-off algorithm for multi-hop wireless networks. The algorithm is developed from two innovative ideas: to reduce the contention in a single flow, it is necessary to reduce the transmission rate of source node; to improve the spatial utility ratio, it is important to reduce the idle slots during transmission. Simulation studies show that the DBA performances better than BEB and AASC for throughput and fairness both in low-load and high-load networks.
  • SUN De-hui; TU Ping; LI Tian-yu
    Computer Engineering. 2008, 34(1): 125-127. https://doi.org/10.3969/j.issn.1000-3428.2008.01.042
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    With the constant development of Internet, the possibility of severe Internet congestion collapse receives much more concern. Efficient congestion control mechanism is one of the key elements to the stability of Internet. To solve the congestion problems, a PID control mechanism based on single nerve cell self-adaptive control algorithm is proposed and analyzed. Simulation that runs on the OPNET software is carried out on multilevel node traffic model. Simulation results show that the proposed mechanism works well in terms of stability of the all levels queue size and system robustness.
  • ZHANG Yi-kui; QIU Xue-song; MENG Luo-ming
    Computer Engineering. 2008, 34(1): 128-130,. https://doi.org/10.3969/j.issn.1000-3428.2008.01.043
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    In order to solve current problems of existing SLA monitor system, the SLA monitor system architecture based on Web services is proposed, the functions of services and SLA monitor process are described. In order to standardize different layers SLA parameters mapping, the general model of SLA parameters mapping is proposed. Implementation of a prototype validates that the architecture can meet multiple services monitor requirements in the environment of next generation Internet.
  • HE Jun; SHU Li; LI Qi
    Computer Engineering. 2008, 34(1): 131-133. https://doi.org/10.3969/j.issn.1000-3428.2008.01.044
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper presents a QoS scheme suit for IPSec security device, and designs a data classification regulation adapt to characteristics of IPSec security device. A principle of queuing model is constructed which takes the delay as the target parameter and the bandwidth as the allocation resource of the queuing system. By this model, an adaptive bandwidth allocation circular schedule algorithm is designed. The results of test show that the IPSec security device guarantees bandwidth and delay for key streams in the situation of heavy burden by QoS mechanism.
  • LI Ni; XIAO Zhen; XU Li-juan; PENG Xiao-yuan
    Computer Engineering. 2008, 34(1): 134-136. https://doi.org/10.3969/j.issn.1000-3428.2008.01.045
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Most of the domestic simulation systems developed based on HLA/RTI at present are built in LAN to form a single federation and realize the connection and interoperability among the federates. In this paper, grid technology is integrated with HLA/RTI to achieve the connection and interoperability among simulation systems in WAN so as to share and reuse the federation-level or system-level function. Also the architecture and key technologies of the connected systems based on grid are introduced.
  • JIANG You-fu; WU Wei-zhi
    Computer Engineering. 2008, 34(1): 137-139. https://doi.org/10.3969/j.issn.1000-3428.2008.01.046
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Wireless Ad Hoc geographic routing has been extensively studied over the last several years. Each intermediate node independently selects the next hop using the given location information of destination. Geographic routing, which eliminates the overhead of route request packet flooding, is scalable and suitable for large scale Ad Hoc networks. However, geographic routing may select the long detour paths when faces the problem of what to do at local minimum node where greedy forwarding fails. This paper proposes a novel geographic routing approach called A* Geographic Ad Hoc Routing (AGAR), which uses the heuristic function and hosts’ location information to optimize this path and overcomes the complexity of the planar graph routing. Simulation results show that AGAR significantly reduces the number of the extended nodes on various network topologies, improves the packet delivery success rate and decreases the average end-to-end delay.
  • Zhu Yi; Zhao Jie
    Computer Engineering. 2008, 34(1): 140-141,. https://doi.org/10.3969/j.issn.1000-3428.2008.01.047
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper introduces the design and implementation of the 802.11 network stack in the 2.6 Linux kernel including the position of the wireless network stack in the Linux kernel, the major functionality, and the communication with both the hardware and user space configuration tools. Since there is much difference between the wired and wireless hardwares, for example, high latency and low reliability, it also performs a code analysis for some important parts of the implementation.
  • Security Technology
  • LI A-li; CHEN Yan-fang; ZHANG Fu-zeng; LI Ling-yun
    Computer Engineering. 2008, 34(1): 142-144. https://doi.org/10.3969/j.issn.1000-3428.2008.01.048
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    The paper provides and realizes the data fusion arithmetic used among the monitor agents in Distributed Intrusion Detection System (DDIS), then testifies that the arithmetic can detect many distributed denial attacks such as SYN flood, Smurf, Land etc in 0.07 s~1 s and take instant countermeasures to block intrusive connections. Based on the analysis of multi-data of many machines, so the arithmetic advances the exactness of intrusion detection largely, overcomes the localization of the traditional detect method of router access control list, and basically realizes detect and alert function without affecting the normal running of the network.
  • ZHANG Jian; YU Xiao-yang; REN Hong-e
    Computer Engineering. 2008, 34(1): 145-146,. https://doi.org/10.3969/j.issn.1000-3428.2008.01.049
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    The initial sensitivity and randomness of the chaotic map make the image encryption algorithm based on chaotic theory become a hot researching field recently. Many people adopt the low-dimension chaotic map to encrypt, however, the high-dimension map is more complex and has more parameters. This paper uses the four-dimension map to encrypt image, at the time of increasing the key quantities, the encryption speed is not reduced. Moreover the abilities of resisting exhaustive attack and statistic attack are enhanced. For an image of 1 024×768 24 bits BMP, encryption cost approximately 0.35 s and decryption cost approximately 0.361 s, then it can meet the transmission requirement.
  • CHEN Li-qing; ZHANG Fu-tai
    Computer Engineering. 2008, 34(1): 147-149. https://doi.org/10.3969/j.issn.1000-3428.2008.01.050
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Designing key management is a difficult task in dynamic secure multicast. This paper proposes a novel key agreement scheme for dynamic secure multicast based on threshold secret sharing. Then the procedures of key agreement of every subgroup and the entire multicast group and updating of group keys with the dynamic change of group members are analyzed in detail. The analysis shows that the scheme is efficient in computation of group members and satisfies the security requirements of key agreement.
  • YU Sheng-chen; WANG Shu; GAO Xiao-yan; HUANG Jiang-lan
    Computer Engineering. 2008, 34(1): 150-152. https://doi.org/10.3969/j.issn.1000-3428.2008.01.051
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    In order to improve the reality, usability and whole performance of network Intrusion Detection System(IDS), this paper presents the approach in which an optimal combination of k(n>k) characters used to classification of IDS is selected from n characters, which is based on genetic algorithms. The optimal set of characters is best for recognizing intruder, which makes classification evaluation target J to achieve maximum. Tests show that this method and IDS(neural network) are useful and available, so as to improve the whole performance of IDS.
  • ZANG Kan; LIN He; TANG Ding; GUO Kai; DING Xiao-lei
    Computer Engineering. 2008, 34(1): 153-155,. https://doi.org/10.3969/j.issn.1000-3428.2008.01.052
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper describes the process of user authentication and key negotiation brought to WLAN by 802.11i, proposes a new mechanism which is designed based on 802.11i standard and 802.1X standard. The mechanism will provide user authentication and key negotiation in access controller under WLAN split MAC structure. Therefore the system will provide robust security services to the WLAN user.
  • ZHAO Mei-ling; ZHANG Shao-wu
    Computer Engineering. 2008, 34(1): 156-157. https://doi.org/10.3969/j.issn.1000-3428.2008.01.053
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper analyzes the security of an elliptic curve cryptography-based combined public key technique due to Tang Wen and some others. Because of some security vulnerabilities of the proposed technique, it gives two kinds of collusion attack methods, and calls the first method as a choice collusion attack. if a customer chooses w customers with some mapping characteristics and colludes them, they will get the 2w-w-1 other customers’ private keys. And it calls the second method a random collusion attack. Two collusive customers first compute their public keys’ difference and , and then they choose public key from public key factor matrix random. Suppose one of the differences between the chosen public key and the two collusive customers’ key is equal to or , the attack succeeds.
  • WANG Yang-qin
    Computer Engineering. 2008, 34(1): 158-160,. https://doi.org/10.3969/j.issn.1000-3428.2008.01.054
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    A Payword-based micropayment protocol is described in this paper. It overcomes Payword’s several shortcomings. In order to make it suitable for m-commerce, only symmetry-encryption is used. Multiple Paywords which reduce customers’ storage and computation can be applied to several vendors. The protocol minimizes the exchange cost of micropayment on the promise of security.
  • ZHANG Xiao-yan; ZHANG Han-ling; LIU Jing
    Computer Engineering. 2008, 34(1): 161-163. https://doi.org/10.3969/j.issn.1000-3428.2008.01.055
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    The Side Match(SM) inevitably causes abnormality in the pixel-value difference histogram, so there are unusual steps in the histogram of pixel-value differences. This provides a clue for steganalysis to reveal the presence of a secret message. To enhance security, a modified scheme is proposed. Classify pixel-value differences |d| into a number of continuous ranges with dynamic, which avoids occurrence of the above-mentioned steps in the pixel-value difference histogram. And adjust d to the value that is closest to the original among all the candidate values, which can effectively reduce the distortion and achieve better imperceptible. Experimental results show the effectiveness of the proposed technique.
  • WANG Guo-jun; WU Bin; WANG Zhi-jian
    Computer Engineering. 2008, 34(1): 164-166,. https://doi.org/10.3969/j.issn.1000-3428.2008.01.056
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    MANETs are vulnerable to many security threats because of dynamic network topology, dynamic membership, unreliable wireless communications and various delays. Existing group rekeying schemes can not be used in MANETs since they seldom consider various delays. This paper proposes the concept of rekeying conflict, and designs a freezing-based group rekeying protocol. In order to achieve rekeying in a consistent way, all cluster heads use the freezing mechanism at the same time to avoid rekeying conflict due to various delays. Simulation studies show that the proposed protocol greatly reduces the failure probability in group rekeying, and that it also reduces the computation overhead in group rekeying.
  • ZHU Jing-jing; ZENG Ping; XIE Kun
    Computer Engineering. 2008, 34(1): 167-169. https://doi.org/10.3969/j.issn.1000-3428.2008.01.057
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    To protect the copyright of large raster maps, a fast robust digital image watermark algorithm is proposed. With the blocking idea, the watermark is embedded into the Fourier semicircle region. When the watermark is extracted, the watermarked image is first preprocessed, and then the spectrum is mapped to standard size with neighbor averaging method, which assures blind extraction, and resolves the problem caused by block-image embedding as well as whole-image extraction. Experimental results show that the algorithm has the advantages of low memory requirements, high efficiency performance and robust for attacks of cropping, rotation, contrast and light variation.
  • LIU Yan; XI Jing; LU Jian-de
    Computer Engineering. 2008, 34(1): 170-172. https://doi.org/10.3969/j.issn.1000-3428.2008.01.058
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper puts forward a method to transfer the work of certificate validation from client to server based on a kind of hybrid trust model supporting cross-certification, and describes the proxy’s general model. According to this hybrid trust model, it presents an algorithm constructing a path base on a depth first search, and summarizes the path validation algorithm. This proxy leads to a real thin client and accelerates the application and development of PKI.
  • WANG Guo-xin; PING Xi-jian; ZHANG Tao; XU Man-kun
    Computer Engineering. 2008, 34(1): 173-174,. https://doi.org/10.3969/j.issn.1000-3428.2008.01.059
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper introduces two steganography algorithms based on image, LSB replace and LSB matching, expounds on these two algorithms’ hiding principle and explains their advantages and disadvantages. It expounds each sort of steganalysis, analyses their detecting principles, gives their applying range and features, and appraises the representative analyzing method of each algorithm. By analyzing the effect on the carrier image of the embedding process, it concludes that the security of the LSB matching is much higher than that of the LSB replace. It also points out the approach and direction to improve the security of information hiding.
  • WU Gu-qing
    Computer Engineering. 2008, 34(1): 175-177,. https://doi.org/10.3969/j.issn.1000-3428.2008.01.060
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper presents a new abnormal traffic detection method: address-aggregated wobble-defended abnormal-traffic control system. This system employs adapted-bloom-filter algorithm to assemble purpose IP addresses, then uses wobble-defended CUSUM algorithm to detect pulsing denial of service attack, the method can work exactly and it declines the spending of system resources at maximum. A detection system employing this method has been run successfully in routers as an individual module.
  • FU Hai-dong; YUAN Xi-guo
    Computer Engineering. 2008, 34(1): 178-180. https://doi.org/10.3969/j.issn.1000-3428.2008.01.061
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    The immune model with low detection efficiency can not adapt to the current network intrusion detection. This paper presents an improved immune model based on double protecting system, combing the biological immune theory with the fuzzy mathematics principle, with two detection processes. According to the Fuzzy Pattern Recognition (F-PR), some useful data samples are condensed, and a rule is abstracted from the special data sample as important evidence for detection. Experimental results show that this model can achieve low false alarm rate and missing rate.
  • LIU Li-qun ; HUO Jiu-yuan; TANG Ding ; WU Wen-zhong
    Computer Engineering. 2008, 34(1): 181-183. https://doi.org/10.3969/j.issn.1000-3428.2008.01.062
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    It’s optimal to implement a RSN in large deployments of secure wireless network. This paper analyzes centralized WLAN characteristics and compares split MAC with other architectures. TKIP encapsulation principles are researched. It gives a new method for split MAC and TKIP implementation scheme in the new centralized WLAN split MAC architecture, and depicts the implementation of software in embedded system, and some further problems to be resolved.
  • LI Cui; LV Lin-tao; LIU Chun-lei; LI Xuan
    Computer Engineering. 2008, 34(1): 184-186. https://doi.org/10.3969/j.issn.1000-3428.2008.01.063
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Based on the analysis and study of distributed cooperation instrusion detection system and its flow dynamic distribution question, a algorithm of dynamic load balancing is proposed. It hashes the destination IP address in packet to map the corresponding packet to scope of the network data collect agent or detection agent’s number, and adjusts the scope according to the performance and load of the network data collect agent or detection agent. Performance analysis and experimental results demonstrate that this algorithm has good performance of dynamic flow and system’s resource is used effectively.
  • Artificial Intelligence and Recognition Technology
  • YANG Wei; LI Jun-shan; HU Shuang-yan; WU Zhi-xuan
    Computer Engineering. 2008, 34(1): 187-189. https://doi.org/10.3969/j.issn.1000-3428.2008.01.064
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Aiming at shortcomings of wavelet transform on compressed image edges, this paper proposes an improved multiscale Wedgelet algorithm based on rate-distortion. Geometry coherence is applied to optimization guideline in process of multiscal Wedgelet decomposition. This paper establishes hidden Markov model, utilizes Hausdorff distance to measure efficiency of image Wedgelet decomposition, natural geometrical property of approximative image edge is obtained. Experimental results show the algorithm is effective.
  • NI Xin-qiang; CHEN Qi; ZHANG Ping
    Computer Engineering. 2008, 34(1): 190-191,. https://doi.org/10.3969/j.issn.1000-3428.2008.01.065
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Affected by speckles, SAR image can not be classified well by using the traditional methods. This paper proposes an automatic SAR image classification algorithm, which extracts the feature based on the gray level co-occurrence matrix, and designs classifier with AdaBoost and error correcting code. Experimental results show that the algorithm is effective for SAR image classification. Compared with maximum likelihood method, the classification accuracy is improved significantly.
  • HUANG Hai-liang
    Computer Engineering. 2008, 34(1): 192-194. https://doi.org/10.3969/j.issn.1000-3428.2008.01.066
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    A case-based reasoning system for mass customization is presented to facilitate the reuse of knowledge or experience that lies in past decision projects. The case’s knowledge representation model for mass customization decision is presented, which covers the case’s information of background, alternatives, solving methods, process and results, etc. The algorithm designed for measuring case similarity is described, which is based on fuzzy mathematics. A software prototype with the functionalities of case retrieval, adaptation, reuse and rules discovering is implemented and verified to be effective and efficient for decision-making under mass customization.
  • XUE Zhao-jun; ZHANG Fan; MING Dong; WAN Bai-kun
    Computer Engineering. 2008, 34(1): 195-197. https://doi.org/10.3969/j.issn.1000-3428.2008.01.067
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    In order to decrease the noises (view, shoe, clothes and etc) and overcome the defects of existing recognition methods, the new technique based on area feature combining Support Vector Machine (SVM) is propseed. Body silhouette sequences of gait are extracted and normalized. The sequences are added together and images of addition are separated into segments. Area features are extracted and used as parameters for gait recognition. Three kernel functions responding to linear, polynomial and Radial Basis Function (RBF) are used in the experiments. This method is applied to USF gait data-set and achieves recognition rate of 82%~100%. SVM recognition is insensitive for noises of view and shoe. Results show robust stability and good performance. Area features present better essential characters of gait. Experimental results show that SVM can function as an efficient gait classifier for recognition.
  • ZHANG Pei; WU Ya-feng
    Computer Engineering. 2008, 34(1): 198-200,. https://doi.org/10.3969/j.issn.1000-3428.2008.01.068
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    The Inverse Compositional Gradient(ICG) algorithm is a novel Inverse Compositional Image Alignment (ICIA) algorithm, which is based on the local orientation. Compared with the ICIA algorithm, the ICG algorithm gives more accurate and reliable matching results under different lighting conditions. The evaluation of the local orientation and gradient is essentially equivalent. Hence four different gradient operators are used to compute the local orientation. White Gaussian noise is added to the input or the template images to imitate those in the real world. Performances of the ICG algorithm using the above four gradient operators under different noise, are compared.
  • LIU Shu; LUO Yu-pin; YANG Shi-yuan
    Computer Engineering. 2008, 34(1): 201-203. https://doi.org/10.3969/j.issn.1000-3428.2008.01.069
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Shape feature is an important visual feature. This paper proposes an approach based on multiscale representation and dynamic programming for the shape matching and recognition. Considering the changes of shape feature points in different scale, multiscale representation of contours is proposed. And the process of contour segmentation can adjust automatically while the amounts of noise and deformation change. It is robust with respect to noise, shape deformation and occlusion. This paper successfullies apply the matching algorithm to the task of pedestrian recognition in the night vision system with good results, also demonstrates the superiority of the proposed method over three traditional shape matching methods.
  • SU Zhao-feng; QIU Hong-ze
    Computer Engineering. 2008, 34(1): 204-206. https://doi.org/10.3969/j.issn.1000-3428.2008.01.070
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Process planning and job-shop schedule are closely related with each other in flexible manufacturing system. The optimality of job-shop scheduling depends on the result of process planning. Symbiotic evolutionary algorithm is used to deal with this problem usually. This paper presents a new definition of individual’s fitness to improve the performance of the algorithm. Simulation results demonstrate the effectiveness of the proposed definition, whose optimization performance is markedly superior to those in the literature and can get much better solutions and cost less time. A new genetic operation is also introduced. Experimental results also indicate the method efficiently improves the performance of the symbiotic evolutionary algorithm.
  • CHEN Qiao-li; WU Huai-yu; CHENG Lei
    Computer Engineering. 2008, 34(1): 207-209. https://doi.org/10.3969/j.issn.1000-3428.2008.01.071
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    A kind of guided crossover for the parameter optimization problem is proposed. The fitness difference between parents would be used for guiding the generation of the offspring so that the Euclidean distance between offspring and better parent will be smaller. There are two reasons for this strategy which can achieve better performance than random crossover. For the continuous functions, firstly, the neighborhood of higher fitness individuals has better individuals, and better offspring can be generated toward the improved direction of fitness. Case studies of the numerical simulations are given to demonstrate that guided crossover has higher efficiency and better ability of global and local search, as compared with conventional arithmetical crossover.
  • XI Li-li; WANG Yi-feng; LIU Rui-hua; ZENG Ping
    Computer Engineering. 2008, 34(1): 210-211,. https://doi.org/10.3969/j.issn.1000-3428.2008.01.072
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Gamut mapping is implemented on the plane of lightness and chroma while keeping hue constant, and color gamut of one device is usually approximated with triangles, which is difficult to reproduce color exactly. This paper analyzes the shapes of device gamut sections, and proposes gamut section boundary descriptor by introducing B-spline curves. A gamut mapping algorithm namely CUSP is given. Experimental results show that the method can increase mapping precision and significantly improve the quality of color image output.
  • LI Shi-jin; CHEN Rong; TIAN Ling; CHEN Yun-hui; ZHANG Yu; JIANG Yong-guang; YU Zhong-hua
    Computer Engineering. 2008, 34(1): 212-214. https://doi.org/10.3969/j.issn.1000-3428.2008.01.073
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Analysis of symptom-syndrome relation in TCM (traditional chinese medicine) is one of the most fundamental and important issues for TCM research, which is of great significance for TCM diagnostics and syndrome research. Data mining technology is adopted to analyze the relation on the available ancient prescriptions. Primary symptoms of the ancient prescriptions are standardized. The relation is mined and the mined knowledge is used to judge primary and secondary syndromes of the prescriptions. The mined knowledge reveals the rules about the relation in TCM. Algorithms based on KNN and Bayesian methods are proposed to mine the relation in TCM. And the comparative experiments show that the correct rate of the algorithm based on Bayesian method achieves 65.76%, and is higher than that of the KNN method whose correct rate is 62.50%.
  • YANG Ying; TANG Ping; WANG Yue-chao; QIU Yan-hang
    Computer Engineering. 2008, 34(1): 215-217. https://doi.org/10.3969/j.issn.1000-3428.2008.01.074
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    An algorithm to optimize artificial neural networks structure based on constructive method is presented. And an improved BP algorithm, LMBP algorithm, is introduced about simplest BP neural network for function approximation. By using the rules of error changing based on quadratic error and gradient reducing, this paper analyzes the optimization of the network structure. By adding hidden neurons or network layers adaptively, a proper structure of the network is got. Simulation experiments are provided to compare the approach with RAN algorithm for solving function approximation. The results show the effectiveness of the algorithm.
  • XU Shi-ming; ZHANG Qiang
    Computer Engineering. 2008, 34(1): 218-220. https://doi.org/10.3969/j.issn.1000-3428.2008.01.075
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    The design of DNA sequence is important in improving the reliability of DNA computation. Proper constrained terms that DNA sequence should satisfy are selected, and then the evaluation formulas of each DNA individual corresponding to the selected constrained terms are proposed. Genetic Algorithm/Particle Swarm Optimization(GA/PSO) Algorithm is presented to solve the multi-objective optimization problem. Comparison of the results with the known DNA sequences is made under different constrained criteria. An example is illustrated to prove the efficiency of the method.
  • JIANG Kai-zhong; LI Zi-cheng; GU Jun-zhong
    Computer Engineering. 2008, 34(1): 221-223. https://doi.org/10.3969/j.issn.1000-3428.2008.01.076
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Automatic abstraction is a key step for information extraction or knowledge acquisition in the aftermath of Information Index. The high quality summarization is significant at the time of massive information. This paper proposes a method by extracting a set of sentences from the text as the abstract based on the semantic structure, which is formed by clustering all the sentences of the document. By employing this semantic structure, summarization quality can be improved. The paper also presents a set of formula to evaluate the automatic abstracting impersonally and efficaciously. Evaluation and experimental results show that the algorithm is effective.
  • SHEN Xiang-jun; CUI Zhi-ming; GONG Sheng-rong; MA Xiao-hu
    Computer Engineering. 2008, 34(1): 224-226,. https://doi.org/10.3969/j.issn.1000-3428.2008.01.077
    Abstract ( ) Download PDF ( )   Knowledge map   Save

    In the research area of Content Based Image Retrieval(CBIR), aiming at extracting the interesting image content to improve retrieval results, this paper proposes an algorithm based on dynamical region pair matching through selecting images by users, uses the Fisher prediction to obtain the dynamical threshold to explore the “interesting” components of the features in the region pair match, adopts those interesting components of the features and the region pairs to retrieve the similar images. The experimental results show the method is effective.

  • ZHANG Shi; DU Kai; ZHANG Wei
    Computer Engineering. 2008, 34(1): 227-229,. https://doi.org/10.3969/j.issn.1000-3428.2008.01.078
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper advances a new optimization algorithm of image registration based on the dynamic combination of genetic algorithm and ant colony algorithm, then it is applied to multimodality medical image registration. This algorithm applied mutual information as the similarity measurement, firstly generates initialization pheromone distribute, then searches the best parameters using genetic algorithm, in the algorithm the strategy of dynamic fusion improves efficiency of searching. Experimental results show that the new registration method can efficiently restrain the local maxima of mutual information function, avoid vast repeated calculations, improve the efficiency of image registration. Furthermore it is excellent in robustness.

  • Engineer Application Technology and Realization
  • ZHU Yue; RONG Meng-tian
    Computer Engineering. 2008, 34(1): 230-232. https://doi.org/10.3969/j.issn.1000-3428.2008.01.079
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper analyzes the architecture of joint TCM decoder and equalizer which suits for IEEE802.3ab std.. 1000Base-T gigabit Ethernet transceiver in 0.13μm CMOS process. Look-ahead technologies available for 1000Base-T are studied to optimize the look-ahead structures employed in the joint TCM decoder and equalizer. Hardware complexity of the decoder is decreased by those approaches without performance scaled down. 8.7k gates of 14tap MA4 decoder is saved after optimization, which is about 9% of original circuits. The gate count of 14tap PDFD is decreased by 9.5k, which saves about 6% in area after optimization.
  • ZHAO Yu-wei; ZHAO Xiao-yu; QIAO Mu
    Computer Engineering. 2008, 34(1): 233-235. https://doi.org/10.3969/j.issn.1000-3428.2008.01.080
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    In order to improve the performance of accessing data, this paper puts all the visited static files, dynamic requests’ response results and the middle objects created while dealing with the dynamic requests into caches. When the same request is visited later, it retrieves the content from the caches and deals with them. It puts forward several all-purpose blueprints about information system caching technology. Experimental results show that the blueprints can markedly enhance the response ability and processing efficiency of application sever.
  • GUAN Gui-xia; ZHU Hong; GUAN Yong; WU Min-hua
    Computer Engineering. 2008, 34(1): 236-237,. https://doi.org/10.3969/j.issn.1000-3428.2008.01.081
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    By analyzing its statistical characteristic, GPS positioning error can be described approximately as two–order Markov process. With the extended Kalman filter introduced, the system status is estimated effectively on the basis of the Singer model of the vehicle. Simulation result shows that the precision of GPS positioning based on extended Kalman filtering is much higher than that based on standard Kalman filtering. So it can be used in eliminating correlative error in low dynamic condition.
  • CHENG Liang-hong
    Computer Engineering. 2008, 34(1): 238-240. https://doi.org/10.3969/j.issn.1000-3428.2008.01.082
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Aiming at the need of the multiple channels and adaptive controlling application in the EMS control system used in metallurgy industry, a design method of multi-channel adaptive controller is proposed. Based on the neural network algorithm, object-oriented method and implemented in assemble language, the multi-channel controller program is compact in configuration and efficient in running, so that the multi-channel controller can implement the task of learning and adaptive controlling, as well as satisfy the need of real time for moving control system.
  • FANG Zhi-bin; SUN Ning-hui; AN Xue-jun; HU Peng;
    Computer Engineering. 2008, 34(1): 241-243. https://doi.org/10.3969/j.issn.1000-3428.2008.01.083
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper analyzes the features of communication in multi-processor chipset and PCI communication. It designs and implements PCI bridge controller in cut-through communication. The frequency of the controller is up to 66 MHz after placing and routing in FPGA, and it can effectively hide the delay of performance brought by burst transfer and delay read in PCI protocol in multi-processor cut-through communication. The PCI controller is used in Godson multi-processor system.
  • WANG Xiao-bin; LIAO Jian-xin; WANG Chun; LIU Zheng; ZHU Xiao-min;
    Computer Engineering. 2008, 34(1): 244-246. https://doi.org/10.3969/j.issn.1000-3428.2008.01.084
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    The limitations of existing test tools are analyzed, and then a new test system——MINSATS is designed. MINSATS is designed to test the signaling procedure of a service provided by mobile intelligent network. Compared with the existing test tools, MINSATS is able to send and receive signaling automatically according to the script which is configured by testing person, check the parameters in a signaling message, check data stored in database, and test all of the signaling procedures that may exist as a batch.
  • REN Hui-feng; YANG Chun-hua; CHENG Peng-sheng
    Computer Engineering. 2008, 34(1): 247-249. https://doi.org/10.3969/j.issn.1000-3428.2008.01.085
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Considering the equipment distribution and various signals existed in building automation systems, a new-style controller based on Rabbit2000 is developeded according to the “Top-Down” scheme, which consists of the modules, such as data acquisition, output control and communication. The controller can detect and control multi-channel input and output independently, and communicate with superior computer by Modbus/TCP. The building controller satisfies the requirements of building automation system and possesses high performance-cost ratio and popularizing value.
  • CAI Yu; YAO Dan-ya
    Computer Engineering. 2008, 34(1): 250-252. https://doi.org/10.3969/j.issn.1000-3428.2008.01.086
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    In complex outdoor environment, there are many limitations when using imaging processing algorithm. This paper proposes an adaptive vehicle extraction algorithm for real-time traffic surveillance system. In order to enhance the ability of adaptation to illumination changes, shadows and system noise in long-term running, the algorithm uses both morphological operation and edge histogram method to extract foreground vehicle objects. And it maintains certain accuracy and processing speed. Experiment on real-world video shows a rather satisfying result. Thus, this offers good foreground information for further vehicle tracking.
  • NAN Zhao-kuo; XU Wen-bo; CHAI Zhi-lei
    Computer Engineering. 2008, 34(1): 253-255. https://doi.org/10.3969/j.issn.1000-3428.2008.01.087
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    With the Java technology widespread applied in the embedded system domain, a Java processor with a 32-bit environment optimized For RTSJ is designed, and is suitable for low power embedded equipments. The processor is implemented by a FPGA chip, which has a new kind of Java stack structure and has a very simple instruction system. It can exeute the Java byte code directly, and provide effective support to the real-time specification for Java. It passes the function simulation in the Xilinx SPARTAN-3 platform, which shows that the Java processor can be implemented in the low cost FPGA chip.
  • CHEN Liang; WANG Shi-jin; ZHOU Bing-hai
    Computer Engineering. 2008, 34(1): 256-258. https://doi.org/10.3969/j.issn.1000-3428.2008.01.088
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Flexible job-shop scheduling problem, containing routing sub-problem and sequencing sub-problem, is an important issue in both fields of combinatorial optimization and production operations. Being an extension of the classical job-shop scheduling problem, it presents additional complexity(NP-hard in strong sense), which makes traditional optimization approaches difficult to achieve an optimal solution in reasonable computational time. In this paper, a filtered-beam-search based heuristic algorithm is proposed to solve the flexible job-shop scheduling problem with the objective of optimal weighted sum of multiple measure performances. The proposed algorithm makes improvement in the generation procedures of branches and the local/global evaluation functions, which makes it possible to solve two sub-problems. the performance of the proposed algorithm is evaluated and compared with those of other representative approaches through simulations, and the results demonstrate the feasibility and efficiency of the proposed algorithm.
  • Developmental Research
  • ZHU Xin-hua; ZHAO Da-zhe; YU Ya-xin; LIU Ji-ren
    Computer Engineering. 2008, 34(1): 259-261. https://doi.org/10.3969/j.issn.1000-3428.2008.01.089
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Some key technologies of platform are studied. Text processing technology is used to extract text features. Image processing technology is used to extract grey, texture and shape features. Text and image processing technology are combined to identify and quantify lesions characteristic in medical image. High-dimensional index is used to organize these features. Semantic Web is used to fuse all kinds of resource together. A prototype system provides users with a high accuracy and high efficiency of the resources retrieval platform to study and communicate with each other. Result shows that the performance of system is highly improved after using above technologies.
  • MA Jiang-lin; ZHAO Zhong-ming; MENG Yu; PENG Ling
    Computer Engineering. 2008, 34(1): 262-264. https://doi.org/10.3969/j.issn.1000-3428.2008.01.090
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper designs and implements a connected component labelling method for remote sensing classification image, which is famous for its massive volume. The problem of huge EMS memory requirement for remote sensing data processing is resolved by an efficient partition scheme. The paper adopts a seed firing algorithm based on queues for connected component labelling in the memory; in order to clean up the label equivalence on the edges of connected blocks, a conflicted table is formed as in the sequential scan labelling algorithm. The method has the following advantage, it can process massive data, requires only one pass over the classification image, and it does not require any re-labelling mechanism to obtain basic statistic information for all connected component region. Experimentas results show that the method is efficient and effective.
  • HAO Ang; SHI Wei-feng; YUAN Jing; ZHAO Rong
    Computer Engineering. 2008, 34(1): 265-266,. https://doi.org/10.3969/j.issn.1000-3428.2008.01.091
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This article discusses the method and process to extract effective information and component form legacy CAD system. Combined with the new development domain of CAD, it proposes a new CAD system model based on software bus architecture and analyses the functions of each component composed the model. With better openess and integrity, the system enables designers at different places to work collaboratively. An example the architecture used in the DrCAD system is presented.
  • YE Xiao-zhou; WANG Jin-lin; ZHANG Jian-dong; DENG Feng
    Computer Engineering. 2008, 34(1): 267-269. https://doi.org/10.3969/j.issn.1000-3428.2008.01.092
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper describes a SIP stack design and implementation for single DSP VoIP ATA. The SIP stack adopts layered structure and session state machine, which accomplishes both basic and extended call functions, and implements 16 bit construction, storage and parse for SIP messages. SIP task is created based on μCosII embedded operating system, it shows good stability and compatibility in practice.
  • SUN Yi-hong; ZHAO Zeng-hui
    Computer Engineering. 2008, 34(1): 270-272,. https://doi.org/10.3969/j.issn.1000-3428.2008.01.093
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper presents an idea to construct a virtual scene system, which can respond with 2D digital map system one another, introduces the details of system’s frame and functions. It puts forward the method used to structure datum of virtual scene, the way to read in datum, how to get the accurate position of observer and how to realize the response function, and gives the program code and pictures. This dynamic response system resolves the limitation of 2D digital map and 3D virtual scene. It also combines the advantage of 2D digital map with the advantage of 3D virtual scene, which overcomes the abstraction of wandering astray of 3D virtual scene.
  • WANG Xu-zong; LIAO Jian-xin; ZHU Xiao-min;
    Computer Engineering. 2008, 34(1): 273-275. https://doi.org/10.3969/j.issn.1000-3428.2008.01.094
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    With the fast development of the mobile communication networks, the need of providing P2P(Peer-to-Peer) services, especially the BitTorrent(BT) file-sharing system in mobile networks is growing. According to the features and actual situation of BT protocols and the 3G networks, a BT file-sharing system architecture based on IMS(IP Multimedia Subsystem) is presented, after analyzing the disadvantages of the extended eDunkey, mobile proxy, and other existing manners, mechanisms of the system and the function of each part are described. The problems of complicated network structure, high cost and hard to promote are solved.
  • JIANG Xiang-gang; ZHOU Xiang-zheng; ZHANG Ji; ZHAO Shi-min
    Computer Engineering. 2008, 34(1): 276-277,. https://doi.org/10.3969/j.issn.1000-3428.2008.01.095
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper makes a further study on system of the gate in urban railway traffic, analyses the requirement of the gate, introduces the current design method and the limitation of software. To mend it, the paper introduces blue print of gate in urban railway traffic based on ReWorks , presents customized process of the operating system and module design of application software.
  • BAI Rui-lin; SHEN Xian-ming; ZHUANG Che
    Computer Engineering. 2008, 34(1): 278-280. https://doi.org/10.3969/j.issn.1000-3428.2008.01.096
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    A software-hardware co-design implementation method of FCMAC based on NiosII is proposed. The method resolves the disfigurement that software implementation of FCMAC has slow speed and hardware implementation needs more resource. It obtains the parameterd of FCMAC according to Matlab simulation, analyzes chiefly the key algorithm which influences operation speed of FCMAC implemented by software in order to get a sound software-hardware partition. Under NiosII IDE development environment, the method implements software modules based on C language and implements hardware modules by user-defined instructions, so the software and hardware modules became unification——the software-hardware co-design is finished. Test result shows that the software-hardware co-design implementation of the FCMAC makes a balance between software implementation of a FCMAC has slow speed and hardware implementation needs more resource. Different software-hardware partition can make interchange between speed and resource.
  • LI Qing-bao; MENG Qing-qian; ZENG Guang-yu
    Computer Engineering. 2008, 34(1): 281-282,. https://doi.org/10.3969/j.issn.1000-3428.2008.01.097
    Abstract ( ) Download PDF ( )   Knowledge map   Save

    Now, as vulnerability scanning can not fully check out complex attack existing in network, many of the traditional security technologies are no longer valid and failure to prevent effective solutions to modern network security problems from information technology, so many methods based on security model are presented. This paper analyzes the PCI extended ROM specification, the mechanism of network communications link switch, the principle of secret information storage, the connection relation of host computers and devices and the characteristic of attack, presents a novel security architecture of network terminal, which is based on extended ROM and PCI bus to carry out the physics gap of a network terminal between interior and exterior network. It discusses the principle and hardware design of the security isolation card based on PCI bus.

  • HONG Xi-jun; CHEN Xiao-feng; ZHANG Ji
    Computer Engineering. 2008, 34(1): 283-284,. https://doi.org/10.3969/j.issn.1000-3428.2008.01.098
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    For making up the deficiency of XML profiles of SCA specification in visually describing waveform component and application assembly or deployment, a concept of visualized assembly and deployment of SCA waveform component are proposed. By graphically represent assembly logic and deployment relationships of waveform components, the architects and developers can communicate design clearly and unambiguously. So some potential errors can be avoided during the process of SCA compatible system design and development. Meanwhile, connections between components are also concluded and summarized. The principles of how to choose proper connection type according to different application scenarios are depicted too. At last an IDE for SCA waveform development based on eclipse architecture is introduced.