Computer Science

Survey on P2P Technology and P2P Traffic Managements

WANG Chun-gang,WANG Hui,YANG Jia-hai

Computer Science. 2010, 37 (4): 1-6.

Abstract

PDF(643KB) ( 1203 )

RelatedCitation | Metrics

P2P applications have become more and more popular in the Internet. They have consumed most of the Internet bandwidth, and affected the performance of traditional Internet applications. In order to solve this problem, rescarchers proposed various technologies to manage P2P traffic, such as P2P traffic control, localization and caching technologies. We first introduced P2P technology and the factors affecting P2P traffic distribution, and then surveyed the latest research efforts on P2P traffic management technologies.

Survey on Search Mechanisms for Unstructured Peer-to-Peer Networks

QIAN Ning,WU Guo-xin

Computer Science. 2010, 37 (4): 7.

Abstract

PDF(377KB) ( 668 )

RelatedCitation | Metrics

The success of file-sharing applications such as Napster and Gnutella makes Peer-to-Peer networks developing rapidly. Efficient resource discovery is the first step towards distributed resource sharing. Recently many search methods have been proposed. We gave an overview of search mechanisms for unstructured Peer-to-Peer networks in this paper. Firstly we approximately classified existing search methods based on the features of them. Then we discussed　critical techniques used in these approaches in detail and compared the performance of them on various metrics, giving emphasis on the success rate, message production, adaptation to dynamic network conditions and maintenance overhead.Finally recommendations were given for future work.

Research of Informationization Processing

WU Zhi-jun,YANG Yi-xian

Computer Science. 2010, 37 (4): 11.

Abstract

PDF(381KB) ( 600 )

RelatedCitation | Metrics

This paper divided informationalization processing into three phases, informationalization construction, information assurance, and indicator. The relationships among the three phases were discussed. The characteristic in each phases was analyzed. The operational protect mechanism for China information assurance based on security baseline policy was presented. The indicator for information assurance was proposed.

Evolvement and Progress of Spatio-temporal Index

ZHANG Lin,TANG Da-quan,ZHANG Chong

Computer Science. 2010, 37 (4): 15-20.

Abstract

PDF(610KB) ( 806 )

RelatedCitation | Metrics

With the development of mobile computing and wireless communication technology in recent years, the spatio-temporal databases have been widely applied in traffic control and global position. A great deal of spatio-temporal indices were proposed for supporting spatio-temporal query and managing massive spatio-temporal objects efficiently.Firstly, summarized the spatio-temporal indices and drew, the evolvement graph, introduced the spatio-temporal query types, then divided the spatio-temporal indices into four types including past, current, future and complete tense according to time, and two types including moving point and moving area according to space, based on the categories expatiated the spatio-temporal indices,finally discussed the directions for further work.

Survey of P2P Semantic Search

WANG Zhi-xiao,ZHANG Da-lu,WANG Yu-hong,WANG Dong

Computer Science. 2010, 37 (4): 21.

Abstract

PDF(519KB) ( 606 )

RelatedCitation | Metrics

A considerable amount of P2P semantic search approaches has been developed. First, the paper introduced a survey schema to review these semantic research approaches from five dimensions,i. e. semantic description method,semantic overlay network structure, query expansion technology, the way of resources clustering and ontology number.Then, the paper compared ten selected typical semantic search approaches by means of the schema. Based on this comparison,we identified and summarized outstanding features of P2P semantic search. Finally, the paper pointed out the necessity to explore new mechanism for resource organization and search,and to focus on a number of open problems,such as P2P semantic search performance evaluation, query expansion technologies improvement and search results ranking.

Study on Global Illumination Using Ray Tracing

ZHANG Quan-gui,WANG Pu,YAN Jian-zhuo,FANG Li-ying

Computer Science. 2010, 37 (4): 27-30.

Abstract

PDF(451KB) ( 1499 )

RelatedCitation | Metrics

Ray tracing is an important method to resolve global illumination. We think that the tracing is the study on the light crossing and efficient sampling.This paper provided a survey on ray tracing algorithms.We introduced the research on real-time ray tracing and GPU based ray tracing, which is the hotspot at present. We pointed out the advantage and disadvantage of ray tracing based global illumination, and the research pivot in the near future. In the end,we prospected the application foreground of ray tracing based global illumination.

Survey on Data Aggregation of Wireless Sensor Networks

KANG Jian,ZUO Xian-zhang,TANG Li-wei,ZHANG Xi-hong,LI Hao

Computer Science. 2010, 37 (4): 31-35.

Abstract

PDF(566KB) ( 1024 )

RelatedCitation | Metrics

Data aggregation is a very important technictue used to reduce the communication overhead and energy expenditure of sensor nodes during the process of data collection in a wireless sensor networks (WSN). The researches on data aggregation of WSN were presented. hhis paper introduced the function and significance of data aggregation,specified the existing and typical classification of data aggregation, pointed out the challenges and the promising research subjects for this research domain, especially compressive sensing.

Fine-grained Distance Controlled Localization Algorithm in Wireless Sensor Networks

ZHANG Song-tao,JIANG Hong-bo,TANG Zhen-hua,LIU Wen-yu

Computer Science. 2010, 37 (4): 36-40.

Abstract

PDF(529KB) ( 551 )

RelatedCitation | Metrics

An improved location discovery algorithm upon previous Hop-terrain localization algorithm was proposed in this paper. At first, the algorithm used distance vector exchange to get hop counts from sensors to anchors and estimate he initial positions of sensor nodes. Then it iteratively updated the sensor nodes' positions by utilizing the initial posilion estimates and information of immediate neighbors of sensors. A fin}grained distance control mechanism was introduced in the position iteration update phase to decrease the position error. I}he distance control mechanism chose the possible position with great accuracy as iteration result based on the constraints of the distance to those neighbors with high weight value calculated by our algorithm. Simulation results show that the fraction of located nodes is increased and the position error is decreased compared with Hop-terrain algorithm.

Reliability Analysis for Flexible Workflow Supporting Exception Handle

ZHANG Jing-le,YANG Yang,ZENG Ming,WANG Yuan-zhuo

Computer Science. 2010, 37 (4): 41.

Abstract

PDF(323KB) ( 566 )

RelatedCitation | Metrics

With growing awareness of the flexible workflow, reliability of flexible workflow system has become more and more important, as one of the main reasons of the impact of work flexible workflow, exception has a great impact on reliability of flexible workflow system. This paper presented a stochastic Petri net based flexible workflow modeling and performance analysis method supporting exception handling，and took e-commerce systems in the logistics system for example, the flexible system's availability, reliability and other performance parameters were analyzed based on the stodrastic Petri net in modeling.

Design and Implementation of Non-intrusive Wireless Sensor Network Tester

LI Tian-lin,HUANGFU Wei,YANG Xu,ZHANG Zhi-qiang,SUN Li-min,LI Fang-min

Computer Science. 2010, 37 (4): 45-48.

Abstract

PDF(450KB) ( 630 )

RelatedCitation | Metrics

Wireless sensor network test tool is great significant for the wireless sensor network's in-depth research and applications. A testing method of non-intrusive wireless sensor network was proposed, and the design and implementation of a wireless sensor network tester based on this method were introduced. I}his tester uses the FPGA to speedily collect the signal of the node inner communication without intrusive to work of the node, and then send the signal to the center scrvicer. With processing these signals, the center servicer can revert the communication of the node or all wireless sensor network. I}he experiment proved that this tester can he very good at collection of the nodes communication data without interference to the work of nodes.

Attack Feature Extraction Using Extension Matrix Theory

CHEN Zhi-xian,HUANG Hao

Computer Science. 2010, 37 (4): 49-51.

Abstract

PDF(326KB) ( 541 )

RelatedCitation | Metrics

With the rapid development of Internet in recent years, computer systems arc facing increased number of security threats. Various soft computing based approaches have been proposed to detect computer network attacks. A method for attack feature extraction based on extension matrix theory was given in this paper. 13y constructing extension matrix on positive and negative examples, the integer programming model for its optimal feature subset selection was built, which will be solved by simple genetic algorithm. Finally optimal rules for detection of specific attack were generated. Experimental results show the achievement of high correct detection rates and acceptable low false positive rates based on benchmark KDI)Cup99 data sets.

Normality of Cryptographic Boolean Functions

WANG Wei-qiong,XIAO Guo-zhen

Computer Science. 2010, 37 (4): 52.

Abstract

PDF(232KB) ( 615 )

RelatedCitation | Metrics

It was pointed out that a good cryptographic 13oolcan function should also be complex enough after it is changed a little. The generalized normality of Boolean functions based on this theory was introduced. The relation between the normality and generalized normality of Boolean functions, and the relation between the generalized normality and the generalized algebraic immunity were proposed. Finally, the normality and algebraic immunity of 13oolcan funclions from the point of algebraic normal form of Boolean functions were presented, which proposed a new way to analyzing the normality and algebraic immunity of Boolean functions.

Differentiated Service in Web Cache Based on Proportional Hit Rate

GAO Ang,MU De-jun,HU Yan-su,PAN Wen-ping

Computer Science. 2010, 37 (4): 55.

Abstract

PDF(309KB) ( 587 )

RelatedCitation | Metrics

Based on feedback control theory, this paper designed a cache space controller through system identification.By dynamically reallocating cache space between different Web objects,the controller could guarantee the priority Web objects a higher hit rate and the proportions of hit rates between different kinds of Web objects constant The experimental results demonstrate the proposed approach achieves the proportional hit rate guarantee under UDSF, LRU and LFU cache replacement algorithms.

Research on Stealth Technology of Windows Kernel-level Rootkits

GONG Guang,LI Zhou-jun,HU Chao-jian,ZOU Yun-ke,LI Zhi-peng

Computer Science. 2010, 37 (4): 59.

Abstract

PDF(379KB) ( 1105 )

RelatedCitation | Metrics

With more and more attention being paid to the Rootkits technology in the fields of cyber-security, various new Anti-rootkits technologies have emerged continually. Under the detection of various Anti-rootkits tools,the conventional Rootkits stealth technology is difficulty to play its role. Based on systematic analysis and research of traditional kernel-level Rootkits stealth technology,this paper presented a three-in-one rootkits stealth technical architecture on the basis of driver module integral transposition,kernel threads injection and IRP inline Hook in depth. Experimental results show that the Rootkits based on this stealth architecture can well bypass the detection of some well-known Anti-rootkits tools (such as Rootkit Unhooker and IceSword) , which fully demonstrates the effectiveness of this three-in-one Rootkits stealth technical architecture.

Leaf-TCAM:A Parallel IP Address Lookup Method and Performance Analysis

ZHU Guo-sheng,YU Shao-hua,DAI Jin-you

Computer Science. 2010, 37 (4): 63.

Abstract

PDF(362KB) ( 671 )

RelatedCitation | Metrics

Features of global routing table and prefix updates were analysized. A parallel IP address lookup scheme based on leaf nodes called Leaf_caf TCAM was proposed. The global routing table is partitioned into subtables and put into K independent TCAM chips. Our scheme can get speedup factor K-1 with0.1*(K-1) redundant. Prefix expansion is not needed and over 90 0 o prefix updates can be done in random mode. Our scheme can partition the routing table evenly and has low cost when the sub-table is overflow. The power cunsumption is only 12 0 0 of traditional single chip scheme.

Clustering Analysis Method Based on Alert Cause

WANG Xiu-li,HAI Mo,ZHU Jian-ming,ZHANG Ning

Computer Science. 2010, 37 (4): 67-70.

Abstract

PDF(451KB) ( 904 )

RelatedCitation | Metrics

The use of intrusion detection has created the problem to investigate a generally large number of alarms. To solve the problem, a clustering analysis method based on alert cause was presented. The correlative alarms with the same attribute were ranged into a clustering according to their causes. The generalized attributes can describe the common characteristic of the alarms. The method can cut down the number of alarms remarkably, simplify the alert analysis, and analyze the security risk in network and application environment accurately. I}herefore, the corresponding measures can be taken in time.

Multi-user Selection with Block Diagonalization Precoding

SHANG Peng,ZHU Guang-xi,SU Gang,TAN Li

Computer Science. 2010, 37 (4): 71.

Abstract

PDF(326KB) ( 682 )

RelatedCitation | Metrics

For the resource allocation and scheduling of downlink multi user multiple input multiple output (MU-MIMO) system, a multi-user proportional fair scheduling scheme based on genetic algorithms (GA) was proposed. By adding some good-gene individuals to the initial population and keeping its gene stable, the convergence of GA was greatly accelerated. Specifically,the base station exploited 131ock Diagonalization (BD) precoding technique to eliminate the inter-user interference. To guarantee the fairness while maintaining the throughput performance, a subset of users was selected to serve at one time slot Moreover, the impact of feedback error on the channel state information was analyzed. Simulation results show that both schemes can achieve a good tradeoff between fairness and throughput with low computational complexity compared with other scheduling schemes.

Security Analysis and Improvement of a Group Signature Scheme

ZHANG Xing-lan

Computer Science. 2010, 37 (4): 75-76.

Abstract

PDF(254KB) ( 692 )

RelatedCitation | Metrics

Recently, Chen, et al. presented a membership revocation algorithm of a group signature scheme, verifiable revocation procedure was invoked to show whether a member is revoked. I3y analyzing Chen's design, a limitation was pointed. At the same time, another improved membership revocation algorithm was proposed. Under the assumption of stronger RSA and Discrete Logarithm problem, the efficiency and the security of the new group signature scheme were proven.

McEliece Public-key Cryptosystem Based on the Maximum F-distance Code

HAN Mu,ZHANG Hong,YE You-pei,XU Chun-geng

Computer Science. 2010, 37 (4): 77.

Abstract

PDF(228KB) ( 832 )

RelatedCitation | Metrics

In terms of F-metric, the maximum F-distance code was constructed, a new modification of the McEhece publie key cryptosystem based on maximum F-distance codes was proposed. The legal party chooses a random matrix as an extra secret key and adds it to the original public key to produce a new modified public key. It makes such cryptosystem effective for resisting the attack based on getting private keys from known public keys. Moreover, using F-metric increases the complexity of the system, making it harder to attack allowing for smaller key sizes. Attacks on such a system were also investigated. It is shown that the McEliece public key cryptosystem based on maximum F-distance codes is security and feasibility.

Research on HRRF Scheduling Strategy Based on TinyOS

LI Hai-cheng

Computer Science. 2010, 37 (4): 80.

Abstract

PDF(241KB) ( 627 )

RelatedCitation | Metrics

Currently the TinyOS task scheduling is based on first comcfirst-served non-preempting strategy, which is not able to give emergency tasks quick response, and throughput of nodes is lower. To address this issue, this paper proposed a preempting algorithm, HRRF ( Highest Response-Ratio First) Algorithm. Through executing real-time tasks within an higher priority, it met system requirement for the real-time response, and improved processor' s response speed. Through executing soft real-time tasks based on Highest Response-Ratio (waited time/run time) scheduling priority strategy,system throughput was increased. The TinyOS test result indicated that under the condition of not affer ting (decreasing) the original performance, HRRF strategy greatly improved sensor network performance in working on real-time tasks.

Load Balancing Algorithm in Structure P2P Systems

WEI Wen-hong,XIANG Fei,WANG Wen-feng,WANG Gao-cai

Computer Science. 2010, 37 (4): 82.

Abstract

PDF(336KB) ( 607 )

RelatedCitation | Metrics

In DHT-based structured P2P system, as a result of the use of distributed hash table, every node can achieve the basic load balancing, but nodes have different capacity to bear the load, and data access has "flash crowds" phenomenon, so there are still load unbalancing in every node. For this disadvantage, we proposed an algorithm for distributed load balancing in which local information adjusts load. And then the problem of single node invalidation is avoided, the implement of algorithm is very simple.

Research on Virtual Tunnel Handover Method in SUPANET

GAO Yu,ZENG Hua-shen,DOU Jun

Computer Science. 2010, 37 (4): 86.

Abstract

PDF(420KB) ( 599 )

RelatedCitation | Metrics

Based on a new architecture called SUPA (Single physical layer User-data transfer & switching Platform Architecture) and the virtual tunnel handover problem which is caused by mobile node when Internet interconnected with SUPA network, this paper proposed two basic virtual tunnel handover methods and a mixed virtual tunnel handover method, then analyzed characteristics of each handover methods. Simulation results show that the mixed method is better than other basic methods.

Multi-granularity Trust Model Based on Individual Expericence

DENG Zhong-jun,WANG Shao-jie,ZHENG Xue-feng,SUO Yan-feng, YU Zhen

Computer Science. 2010, 37 (4): 91.

Abstract

PDF(440KB) ( 543 )

RelatedCitation | Metrics

In distributed networks, the trust evaluation for the same quality services of one node conducted by different access nodes is different. On the one side,it has relation to the direct mutual experience of the accessing node. And on the other side, it has relation to the accessing node's interest and opinion about the service evaluation. Some nodes think a lot of the download speed, while some other nodes put more emphasis on the security and reliability of the service, which cannot but lead to influence the veracity of trust evaluation. In order to eliminate the infection resulted from the trust evaluation difference of individual nodes, a multi-granularity trust model based on individual experience was put forward. In this model, experience factor and multi trust information recording methods were employed. Experiment and analysis show that the proposed model has better improvement in granularity and veracity of trust evaluation.

Multi-stage Election Based Non-even Cluster Heads Selection Algorithm for Wireless Sensor Networks

WANG Lei-chun,MA Chuan-xiang

Computer Science. 2010, 37 (4): 95.

Abstract

PDF(341KB) ( 573 )

RelatedCitation | Metrics

In order to save and balance energy consumption of cluster heads (simply CHs) , a multi-stage election based non-even cluster heads selection algorithm for wireless sensor networks, MSEI3NECHSA, was proposed. Firstly, MSEI3NECHSA built energy distribution graph of candidate CHs by thre}stage election mechanism including spatial virtual grids,CHs and Sink. 13y adopting simulation annealing algorithm based on the parameters including energy of candidate CHs and distance among them, MSEI3NECHSA then made CHs' competition range vary with the distance from CHs to Sink. Simulation results show that MSEI3NECHSA can cluster CHs by less energy and balance energy consumption of CHs in WSN.

Anycast Routing Protocol in Wireless Sensor Networks Based on Link Quality

XU Xin,GU Yun-li, QIAN Huan-yan

Computer Science. 2010, 37 (4): 99.

Abstract

PDF(412KB) ( 565 )

RelatedCitation | Metrics

In Wireless Sensor Nctworks(WSN)，there is the probability of transmission failure and energy waste for retransmission because of link quality problem. Routing criteria of conventional anycast routing protocols are based only on routing distance but not in consideration of link quality problem, it will lead to unreasonable routing weight setting if we apply those protocols in WSN. A WSN anycast routing protocol based on link quality was proposed. Calculation of link duality was discussed. The routing criterion of our protocol is based on the hybrid of load balancing and minimum total energy cost,and we balanced the two factors by parameter a. In contrast to conventional anycat routing protocols,experiment results show that performances in WSN are improved in terms of energy balancing and network energy consuming hence network lifetime is optimized.

Unified Framework for Block Cipher Resembling AES Algorithm and its FPGA Implementation

YANG Hong-zh,HAN Wen-bao,DONG Bo

Computer Science. 2010, 37 (4): 103.

Abstract

PDF(250KB) ( 733 )

RelatedCitation | Metrics

AES algorithm was modularized and the operation of it was generalized. A unified framework was proposed for block cipher resembling AES algorithm. In this framework, not only AES encryption or decryption can be realized,but also the specific block cipher algorithm can be set through external parameters. I}he use of cryptographic algorithms is more flexible and security. FPGA realization for the unified framework was also given. The testing results show the scheme is feasible and has a higher speed.

Study and Design of Adaptive Network Coding MPTCP in Wireless Mesh Networks

XIA Zhuo-qun,CHEN Zhi-gang,ZHAO Ming,LIU Jia-qi

Computer Science. 2010, 37 (4): 106.

Abstract

PDF(459KB) ( 681 )

RelatedCitation | Metrics

It is a challenge that network coding achieves wide use in wireless mesh networks. Network coding must be compatible or friendly with TCP in practice. We proposed a mechanism that naturally adds network coding to current network systems. In our scheme, according to destination feedback degree of freedom of encoding block, the source node modifies the coded data block and delays of packet delivery minimize reordering and timeouts. Using multipath routing,the source side can send more encode packets and the transmission rate becomes more faster, as a result, the network achieves higher throughput compared to TCP in the presence of lossy wireless links. Comprehensive simulations and resups verify the validation of the theoretical results in the paper.

Multi-layer Centerless Hierarchical Structure in Ad Hoc Networks

WU Jing,GU Guo-chang,HOU Guo-zhao,SUN Ting

Computer Science. 2010, 37 (4): 110.

Abstract

PDF(326KB) ( 542 )

RelatedCitation | Metrics

Based on the onclaycr hierarchical structure, a multi-layer hierarchical structure was introduced to solve the scalability problem in ad hoc networks better. Aiming at the problems in this structure such as poor structure stability and large structure maintenance overheads,a multi layer centerless hierarchical structure was proposed. It adopts a centerless scheme, and the structure depends on some relatively stable clusters, which will make the structure stabler. To reduce structure maintenance overheads, a new addressing scheme was adopted, which can avoid the frequent alterations of a node' s hierarchical ID caused by a clusterhead' s change. Both theoretical analysis and simulative analysis were made to compare the two structures' performance. hhe conclusion shows that multi-layer centerless hierarchical struclure outperforms the current multi-layer hierarchical structure on structure stability and structure maintenance overheads.

Research of Grid Selective Resources

LI Xiang,SUN Hua-zhi

Computer Science. 2010, 37 (4): 114.

Abstract

PDF(245KB) ( 550 )

RelatedCitation | Metrics

As such a complex distributed grid computing system, this system wants to study the problems of the multiconstrained task expectance under the condition of limited capacity. So it behaves greatly important. Based on the unrcasondable uesing,we adressed the selective configuration of grid resources. This article analyzed the grid environment, data sharing and task scheduling, studing under the conditions of the limited capacity of the multi-constrained task expcctancc，to adress the selective configuration of grid resources. Considering the impact of multi tasking and resource constrained circumstances,a pulse function mathematical model was introduced and simulated to verify the feasibility of the model.

Research of Software Project Risk Management Based on Rough Set Theory

DING Jian-jie,HAO Ke-gang,HOU Hong,GUO Xiao-qun

Computer Science. 2010, 37 (4): 117.

Abstract

PDF(250KB) ( 698 )

RelatedCitation | Metrics

The complexity of software process leads to that risk data is uncertain and rough. Based on the rough set theory,this paper analyzed history data of software risk management. Using concept of ectuivalence class and discernible matrix, put forward risk factors of correlation analysis system, and defined calculate formula of dependence coefficient and risk factor discernible matrix. I}he work provides policy to risk control and help organization to find their shortcomings in risk management and improve their process.

Vulnerability Analysis Framework for Binaries Based on Model Checking

WANG Chun-lei,LIU Qiang,ZHAO Gang,DAI Yi-qi

Computer Science. 2010, 37 (4): 120.

Abstract

PDF(457KB) ( 712 )

RelatedCitation | Metrics

In order to analyze vulnerabilities in executable programs, a vulnerability analysis framework for binaries based upon model checking was proposed. Firstly, the abstract model of binary was defined, and the formal models of vulnerabilities based upon finite state automaton and the representations of software security attributes based upon event system were described. Then, the model checking based vulnerability analysis process and algorithm were proposed with respect to the abstract models of binaries and the security attributes to be checked. After that, the prototype of vulnerability analysis tool was designed and implemented based upon the framework. I}he illustrative sample program was analyzed to show in detail the principles of the framework, and the experimental results show the effectiveness of the analysis method.

Internet-based Distributed Software Production Line Framework

WANG Xiao-bin,GUO Chang-guo,WANG Huai-min

Computer Science. 2010, 37 (4): 125.

Abstract

PDF(329KB) ( 579 )

RelatedCitation | Metrics

This paper presented DSPLF (Distributed Software Production Line Framework)，a framework for developing distributed software production lines on Internet. DSPLF proposed an architecture for distributed software produclion line, including the composing components and their relationships, based on this architecture, methods and detail processes for constructing and running distributed software production lines were given, also with a distributed software production line project for monitoring,which was developed based on this framework.

Web Service Composition Architecture Based on Life Cycle

YANG Dan,SHEN De-rong

Computer Science. 2010, 37 (4): 128.

Abstract

PDF(347KB) ( 553 )

RelatedCitation | Metrics

Web service composition makes it possible for Web service reused, assembled and providing value-added service .It is important to propose a complete,validate architecture for development and research of Web service composition. A Web service composition architecture based on the life cycle was proposed. The architecture supports process modeling, composition and execution. At the phase of modeling a GUI tool was provided to define the process which then was parsed by document parser to be stored in the service storage. In the composition phase services matcher based on ontology was introduced to solve service heterogeneous problem, and QoS of composite service was also considered.In the execution phase process monitor, exception handler and transaction adapter were provided.

Multi-keywords Query Expansion with OLCA Based Concept Tree Pruning

WANG Zhao-long,LI Xia,XU Rui-fang

Computer Science. 2010, 37 (4): 132.

Abstract

PDF(447KB) ( 602 )

RelatedCitation | Metrics

In semantirbased query expansion, computing expansion words and its weight is a key step to describe the needed query. We proposed a method called OLCA (keyword to concept method) , the idea comes from LCA(Local context analysis). We made some improvement and applied it to multi-keywords ctuery with different weight according to their attribute to the query. Combined with concept tree based on WordNet, we made the query expansion performed from both semantic and the real ctuery documents aspects, and calculated the weight of expansion term based on this technique. Compared with the one based on semantic or the traditional expansion which merely is based on statistic, the experiments reveal that this method can achieve a better query quality.

Research and Design of Software Component Description Language Based on the Reconfiguration Router Unit

LIU Qiang,WANG Bin-qiang,YUAN Qin-hua

Computer Science. 2010, 37 (4): 136.

Abstract

PDF(249KB) ( 580 )

RelatedCitation | Metrics

With the research on muter from the improving perfermence to fuction extention, the reconfiguration muter becomes a good sample in the muter function-extend field. The design of component of reconfiguration muter imported the thinking in software-component Based on the development of software component model, the paper gave the research on the description and the management of reconfiguration muter component model. hhe research and design of description language and description method of software component have also laid a solid foundation for the next phase of development the muter.

Research on Logging Mechanism of Self-healing Database System

XIE Mei-yi,ZHU Hong,FFNG Yu-cai

Computer Science. 2010, 37 (4): 139.

Abstract

PDF(427KB) ( 583 )

RelatedCitation | Metrics

Selective recovery allows a self-healing database system to undo only those operations affected by malicious transactions instead of rollbacking the whole history after malicious attacks, but this function relies on the logging mechanism's support on inter-transaction dependency tracking and longtime preservation of before image data. The inadequacy of traditional logging mechanisms and the problems of the methods used in current prototypes were analyzed.Following this, a new logging structure was proposed which contains inter-transaction dependency information and replaces the rollback segments in traditional logging mechanisms with the before image tables. Database recovery approaches based on the new mechanism were also presented. At last, we analyzed the performance and space overhead of our method, and compared it with other methods.

Research and Implementation of Parallel Programming Environment for Hybrid Parallel Computing System

WU Hua-bei,SUN Ji-zhou,WANG Wen-yi

Computer Science. 2010, 37 (4): 143.

Abstract

PDF(360KB) ( 640 )

RelatedCitation | Metrics

This paper analyzed the parallel processing pattern in the hybrid parallel system of CMP-Cluster and implemented a parallel programming environment based on hierarchy top-down approach. This system can simplify the parallel programming and improve the efficiency and correctness.

Algorithm for Increment Update of k-Anonymized Dataset

SONG Jin-ling,ZHAO Wei,LIU Xin,HUANG Li-ming,LI Jin-cai,LIU Guo-hua

Computer Science. 2010, 37 (4): 146.

Abstract

PDF(499KB) ( 619 )

RelatedCitation | Metrics

K-anonymity is an effective method to prevent linking attack and protect privacy. The main idea of k-anonymity is generalizing the values on a set of special attributes named ctuasi-identifier, so that gains a k-anonymized dataset in which the values of each tuple on quasi-identifier must repeat at least k occurrences. Although k-anonymized dataset guarantees privacy, the k-anonymized dataset needs to be updated constantly because the original dataset updates occasionally after a version of k-anonymized dataset has been existed. So, how to update the k-anonymized dataset as well as the original dataset becomes an urgent problem. To solve this problem, based on the detailed analysis to various update situations of the k-anonymized dataset, the increment update algorithms for the k-anonymized dataset were presented.

Software Behavior Model Based on System Calls

TAO Fen,YIN Zhi-yi,FU Jian-ming

Computer Science. 2010, 37 (4): 151.

Abstract

PDF(674KB) ( 1165 )

RelatedCitation | Metrics

Modeling program behavior based on system call has become the hot topic in intrusion detection since system call can reflect the program behavior in some degree. This paper studied three different types of modeling methods that are dynamically modeling, statically modeling and hybridly modeling as the breakthrough point, and concluded that the development process of behavior models can be divided into three stages: initial stage, developmental stage and synthetical stage. I}he evaluation and comparison experiments were done to find the inherent relations and development track of some typical models in different stages. The whole analysis in this paper indicates that the future trend of behavior modeling methods is to develop a software behavior model with high detection capability, completeness, and actual feasibility through the combination consideration of the static techniques with dynamic techniques, the control flow with data flow,and the other real-time information such as environment variables and context information.

Research on Architectures and Key Techniques of the Collaborative Platform on Clustering Applications for Governments

ZHANG Xue-wang,LI Zhou-jun,SHEN Wei

Computer Science. 2010, 37 (4): 158.

Abstract

PDF(494KB) ( 651 )

RelatedCitation | Metrics

The collaborative platform on clustering applications for governments consists of six largcscale systems. The application and technique architectures of the collaborative platform were elaborated in this paper, and the major key technologies on the platform were also expounded, which include realization of many governments applications' scale integration and collaborative application, business model driven software development platform based on SOA, uniform resource management, and security improvement of Web services. The platform having been running shows that rural distrios and counties can build their own platforms based on the provincial platform to maximize the utilization of hardware, software resources and administrative resources of the provincial government collaborative platform; thus the planning of urban and rural e-governments' construction and maintenance is achieved, efficiencies of government affairs' administration and collaborative work are improved.

Study on Software Aging Forecasting in Markov-Model-based MBP Neural Network Methods

LIN Yi-jie,ZHANG Wei-qun,ZHOU Min,LAI Qing

Computer Science. 2010, 37 (4): 163.

Abstract

PDF(312KB) ( 687 )

RelatedCitation | Metrics

Because of the degeneration of system environment, running software system doesn't avoid software aging.This paper gave a reforming MBP Neutral Network(MBPNN)to represent the Markov Model(MM). This method can be used to supply the gap of the inaccuracy of forecasting result, when only using Markov model. According as the result, software rejuvenation can be implemented reasonably. Study results indicate that the method can forecast the status of running software system efficiently.

Impreciseness:Meaning and Property

HUANG Ying-hui,LI Guan-yu

Computer Science. 2010, 37 (4): 167.

Abstract

PDF(378KB) ( 587 )

RelatedCitation | Metrics

For a long time the randomness and the fuzziness are considered as the two aspects for explaining the uncerlain phenomena, but the roughness being introduced later promotes to form the triune framework for describing the uncertainty.The randomness is an objective attribute, but the fuzziness and the roughness arc related to human's cognitive activities so that are collectively known as the impreciseness. The impreciseness becomes a promising research direction for the uncertain phenomena. The impreciseness is the cost and defects which human have to bear when cognizing the uncertain objects,in detail,the fuzziness is the loss of clear understanding for a cognitive object in order to obtain the cognitive possibility and efficiency; the roughness is the lack of exact description for a cognitive object as to be explained approximativcly with two subsets of the existing knowledge set. The fuzziness is derived from the human's classifying cognitive objects, and the roughness from the imperfection of human's existing knowledge. The impreciseness has three primary properties; subjectivity, classifying-based, and relying on existing knowledge. To decrease the impreciseness should start with the existing knowledge, increasing the cauantity and improving the orderliness.

Application of Non-negative Matrix Factorization in Tag Semantics Analysis

ZHANG Lei-ming,LI Qiu-dan,LIAO Sheng-cai

Computer Science. 2010, 37 (4): 171.

Abstract

PDF(434KB) ( 585 )

RelatedCitation | Metrics

With the development of Web2. 0 technologies, social tagging systems are becoming more and more popular,which makes tags widely used to retrieve, categorize, and manage users' collections. However, people arc free and uncontrollable to use tags, resulting in a large number of tags that are redundant, unclear in semantics. To deal with this problem, we proposed a tag semantics mining algorithm based on non-negative matrix factorization method. We got a tag subspace containing a series of semantic related tag-bases by factorizing tagged data of users using non-negativity constraints,to make synonymous and related tags into the same tag-basis,and categorize polysemous tags into different semantic tag-bases. Simultaneously, the tasks of grouping synonymous tags and distinguishing polysemous tags were done by the proposed approach. A large number of experiments demonstrate the effectiveness of the proposed algorithm on mining tog semantics.

Convex Combination Track-to-track Fusion Algorithms Based on Piecewise RTS Smoothing

CHEN Jin-guang,GAO Xin-bo

Computer Science. 2010, 37 (4): 175.

Abstract

PDF(365KB) ( 1005 )

RelatedCitation | Metrics

Aimed at the problem that the performance of convex combination track-to-track fusion algorithm degrades in the case that the process noise of dynamic system does not equal zero, RhS smoother was introduced to improve the performance. hhe traditional RhS smoothing algorithm can be work when all the filtering data arc obtained, and the outputis delayed largely. Piecewise RTS smoothing algorithm was presented to solve the problem. On the one hand, track-to-track performance can be enhanced; on the other hand, real-time ability can be remained. Furthermore, aimed at the calculation ability of local node owns or not, Smoothing First and Fusing Next (SFFN) and Fusing First and Smoothing Next (FFSN) algorithms based on the piecewise RTS smoothing algorithm were presented according to the chance of implementing the smoothing process. The new methods exceed the performance of the traditional convex combination fusion algorithm and the optimal fusion algorithm in a wide range of process noise. Simulation results show the new algorithms' validity and superiority.

Integrated Interference Matrix Based Ant Colony Algorithm for Assembly Sequence Planning

PENG Tao,LI Shi-qi,WANG Jun-feng,FANG Jian-xin

Computer Science. 2010, 37 (4): 179.

Abstract

PDF(387KB) ( 657 )

RelatedCitation | Metrics

An ant colony algorithm was constructed to search the best optical assembly sequence rapidly in an assembly sequence planning. An assembly model for expressing the touching and preferential relationship among the assembly parts effectively was built by using directed assembly graph. On the basis of discussing the form and property of the assembly interference matrix,a concept of integrated interference matrix and its transformational methods was proposed.The integrated interference matrix and the tool list matrix can express the basic assembling information fully. The derivation of assembly feasible direction calculation based on the integrated interference matrix, the state transition probability of assembly parts,and the update rules of algorithm information were analyzed simultaneously. According to the changing frequency of feasible direction and the changing time of the assembly tools, the optimized functions for ant colony algorithm were then established and two examples were also proposed to prove the practicability of the algorithm.

Automatic Analysis of Chunk-extension Sentences

WEI Xiang-feng,ZHANG Quan

Computer Science. 2010, 37 (4): 183.

Abstract

PDF(427KB) ( 649 )

RelatedCitation | Metrics

The chunk-extension sentence is a kind of sentence, which is defined with specific conceptual knowledge in the theory of hierarchical network of concepts. This paper introduced some semantic categories of chunk-extension senfences and typical head verbs in them. Through these typical verbs, the eigen semantic chunk and semantic category of a chunk-extension sentence can be activated. After testing with the conceptual knowledge of chunk-extension semantic categories,the semantic structure of a sentence can be revealed. An experiment to automatically analyze some chunk-extension sentences from corpus was done based on a system which is designed for analyzing semantic categories of senfences. hhe precision of experimental result is 71. 29%. The errors arc mainly caused by recognizing head verbs and multi semanti}category of a verb. To analyze chunk-extension sentences well will be help to tackle the puzzle of multiverb in processing Chinese sentences.

Discrete Particle Swarm Optimization Algorithm for Vehicle Routing Problems

WEI Ming,JIN Wen-zhou

Computer Science. 2010, 37 (4): 187.

Abstract

PDF(419KB) ( 967 )

RelatedCitation | Metrics

A fuzzy programming model was built to optimize total cost of vehicle routing problem, where vehicle travel time and customer service time were fuzzy. The mode was firstly converted into a deterministic one, and then it was solved by a discrete particle swarm optimization algorithm which redefined the equation of particle motion and algorithms of discrete variables and designed exclusion operator to maintain the population diversity. In comparison with both standard genetic algorithm and standard particle swarm optimization algorithm, it can effectively avoid the search being trapped into local optimum and achieve satisfactory results.

Classical Optimal Planning Method Based on Propositional Satisfiability

LU Shuai,LIU Lei,JIANG Hong,WEI Wei

Computer Science. 2010, 37 (4): 192.

Abstract

PDF(451KB) ( 597 )

RelatedCitation | Metrics

Graphplan-based encoding is a novel encoding adopted by SATPLAN2006,which is the statcof-thcart optimal planner of International Planning Competition in 2006. It firstly defined the concepts and characters about encoding.Based on Graphplan-based encoding, it proposed a new encoding, called FA-based encoding, by reducing the mutex axioms of actions and appending frame axioms. It justifies the soundness and completeness of the corresponding proposed encoding. We designed and implemented the corresponding planner called FA-SP, and tested it with benchmarks adopted by International Planning Competition, respectively. Comparing with SATPLAN2006，the results validate th at for both all two domains tested, the scales of encodings of FA-SP are compacted and solver efficiencies of them arc improved, and it is notes to sec that for sequential planning problems Blocks World, the scales of encodings of FA-SP are compacted nearly 40% and solver efficiencies of them are improved nearly 2 times; for parallel planning problems Logistics, the scales of encodings of FA-SP arc compacted more than 75%. The solver efficiencies of FA-SP are improved several times, and the scales of them are enlarged less than 5% for overwhelming majority problems.

Face Gender Classification Based on Independent Component Analysis and Genetic Algorithm

WANG Zhen-hua,MU Zhi-chun

Computer Science. 2010, 37 (4): 197.

Abstract

PDF(430KB) ( 564 )

RelatedCitation | Metrics

The gender of a face is almost its most salient feature, and realizing automatic gender classification according to the face image will boost the performance of face retrieval and face recognition in large face database. This paper proposed a new gender classification method combining independent analysis (ICA) and genetic algorithm(GA). The Fast ICA algorithm was used to derive independent basic images and projection vectors out of the face images and each image was represented as a feature vector projected in the low-dimensional space spanned by the basis vectors. Then,a genetic algorithm was used to select a subset of features which scan to encode important information about gender form the low-dimensional representation. Finally, the SVM classifier was trained to perform gender classification using the selected independent features subset, The local features extraction of ICA, the features selection ability of UA, and the strong classify ability of SVM were combined reasonably. The experiment results show that the method gets a better classifier performance.

Hybrid Systems Model and Optimization Control of Aerial Robotics

LI Jian-qiang,PEI Hai-long,LU Ya-hui

Computer Science. 2010, 37 (4): 201.

Abstract

PDF(336KB) ( 630 )

RelatedCitation | Metrics

To solve the optimization control problem of aerial robotics (small unmanned helicopter) nonlinear system with six degree of freedom(DOF) , this paper proposed a method that models and lower-orders based on hybrid systems,obtains the maximal controlled invariant sets by states constraint, then solves the optimization problem by mixed integer quadratic programming in the sets. Finally, a model is modeled and simulated, the result proves the validity and efficiency.

Attribute Reduction of Set-valued Information Systems Based on Neighborhood Relation

CHEN Zi-chun,QIN Ke-yun

Computer Science. 2010, 37 (4): 205.

Abstract

PDF(332KB) ( 540 )

RelatedCitation | Metrics

Set-valued information systems, in which some objects may have more than one value for an attribute, are generalized complete information systems. Introducing a neighborhood relation to set valued information systems,we established a rough set approach based on the neighborhood of every object as elementary set. Furthermore, in order to simplify knowledge representation, we proposed the concept of neighborhood consistent set and neighborhood reduction,and discussed the properties and judgments of neighborhood consistent set. Finally, we obtained the approach to compute neighborhood reduction for set valued information systems by the discernibility function.

Research on Multi-Agent Service Recommendation Mechanism Based on Bayesian Network

DU Jing,YE Jian,SHI Hong-zhou,HE Zhe,ZHU Zhen-min

Computer Science. 2010, 37 (4): 208.

Abstract

PDF(475KB) ( 699 )

RelatedCitation | Metrics

In ubictuitous computing environment, services recommendation needs to satisfy the system's mobility and heterogeneity. A multi Agent service recommendation mechanism was put forward, bayesian network and clustering method and update mechanism were used. It is fully taken into account that contexts affect the result of recommendation. The system consists of Agents which accomplish the functions such as history context collecting, knowledge training, poficymaking recommendation and self-learning. Furthermore, by designing the information of communication, flow control and data shared-path among Agents were constructed.

Power Load Forecasting Model Based on Knowledge Discovery

DOU Quan-sheng,SHI Zhong-zhi,YU Er-keng,YANG Bin,LIU Zhong-yao

Computer Science. 2010, 37 (4): 212.

Abstract

PDF(329KB) ( 511 )

RelatedCitation | Metrics

Load forecasting is a traditional research field of power system, this work made an analysis for meteorological and economic characteristics of Heilongjiang Province,and put forward the load forecasting model based on knowledge discovery. First of all, generate the initial prediction curve by traditional method, and extract related rules from the historical meteorological data and load data,update the initial prediction curve using these rules,to generate the final prediction curve. The model has been used in software system of load forecasting of Heilongjiang power grid Co. , Ltd, obtained desired results.

Chinese-Uyhur Sentence Alignment Based on Hybrid Strategy

TIAN Sheng-wei,TURGUN Ibrahim,YU Long,JAMILA Wushouer,YANG Fei-yu

Computer Science. 2010, 37 (4): 215.

Abstract

PDF(409KB) ( 616 )

RelatedCitation | Metrics

This paper proposed a hybrid algorithm of sentence alignment in Chinese-Uyhur parallel corpora. Aiming at the shortcoming of mistake spread in alignment algorithm based on length, this paper presented a new kind of suppression strategy for mistake spread. By using csentence length and ChinescUyhur correspondence information, the anchor points with 1:1 pattern sentence pairs are identify to suppress mistakes spread. Among anchor points,a approach based on both length and punctuation is used to align sentences. Experimental results verify the high precision of identifying anchor points and the effective restraint of the spread of mistakes; Hybrid alignmentd algorithm avoids the weakness of high time complexity algorithms based on words. In addition, its performance is improved more compare with traditional alignment algorithms, and increase alignment accuarcy from 95. 0% to 97. 6% and recall from 96. 8% to 98. 2% , and the validity evaluation method can find the noised alignment efficently.

System Decision Making Method Based on Vague Bidirectional Approximate Reasoning

ZHANG Qian-sheng,JIANG Sheng-yi

Computer Science. 2010, 37 (4): 219.

Abstract

PDF(365KB) ( 519 )

RelatedCitation | Metrics

A similarity degree of vague value was firstly proposed, then a new weighted similarity measure of vague sets was given based on OWA operator, in which the weights were produced according to the requirement of decision maker.Finally, a simple bidirectional vague reasoning method for vague decision system was put forward on the basis of similarity measure between vague fact and the antecedent (or consequent)portion of vague rule. The presented method is very flexible and feasible, which provides a powerful tool for dealing with vague decision making problems in complex enviroment such as industry inspection and medical diagnosis systems.

Data Clean Method Based on Bus Model

YANG Meng-ning,ZHAO Peng,ZHANG Xiao-hong,LI Peng

Computer Science. 2010, 37 (4): 224.

Abstract

PDF(242KB) ( 585 )

RelatedCitation | Metrics

Data cleansing is an important part for ensuring data quality. The existing cleaning methods are often too dependent on a specific application, can not be reused. In order to improve the reusability and scalability of the clean meth od, a data clean framework was build which is based on bus model and reusable. The data clean tool which has independent clean function is registered on the bus through the adapter. The clean function is finished by calling the clean components which is registered on the bus. Finally, how the method works in the really scene was described. The method was proved has good value of application.

DSmT-based Mobile Robot Map Building and Sensor Management

YANG Jin-yuan,HUANG Xin-han,LI Peng

Computer Science. 2010, 37 (4): 227.

Abstract

PDF(330KB) ( 568 )

RelatedCitation | Metrics

A new information fusion method namely DSmT (Dezert Smarandache Theory) was introduced to solve the problem of robot map building in an unknown environment. "hhe grid map method was adopted, and according to sonar sensor mathematical model under DSmT framework, a group of general basic belief assignment functions (gbbaf) was constructed based on classical DSm model to deal with the uncertain and imprecise, and even high conflicting information in the unknown environment. And a simple but very effectual sensor management method was proposed to completely climinate the multi-reflection and crosstalk of the sound wave in the complex environment At last,Pioneer 2-DX mobile robot was used to carry out experiments of map building with DSmT and DST (Dempster-Shafer Theory). The corrclative 2D general basic belief assignment (gbba) map was constructured. The comparison between the two results testifies the validity of DSmT and the proposed sensor management method in unknown environment It supplies a powerful theoretie evidence for fusing dynamic high conflicting information.

New Algorithm for Discretization Based on Information Entropy

YUE Hai-liang,YAN De-qin

Computer Science. 2010, 37 (4): 231.

Abstract

PDF(333KB) ( 600 )

RelatedCitation | Metrics

The discretization of continues attributes is always with great contribution to the followed process of machine learning or data mining. A new algorithm based on information entropy for discretization of decision table was proposed. Through inconsistency checking of decision table, we deleted some redundant cut points on the basis of preliminary discretization scheme. The experiments of classification of discreted data were performed by using SVM, and meanwhile compared with other algorithms, the presented algorithm is effective.

Seismic Exploration Attribute Optimization Based on Multi-Agent Parallel Genetic Algorithm

LIU Qi-cheng,ZHENG Wei-min

Computer Science. 2010, 37 (4): 234.

Abstract

PDF(363KB) ( 544 )

RelatedCitation | Metrics

The method of design and implementation for parallel genetic algorithm is based on thick grain, with the multi-Agent combined with genetic evolution technology, to optimize the seismic attribution selection. A multi-Agent system includes two kinds of Agents:件叔ent and MAgent, they can exchange the useful information which can represent the current situation of evolution. The methods arc benefit to improve the performance of parallel genetic algorithm, and to raise the searching efficiency of the parallel genetic algorithm. The projecting feature in the parallel module is less communicating overhead. It was used in the optimization of oil exploration attribute selection and the result is good.

Fast Algorithm of Dynamic Shortest Paths Based on Discrete Varying-weight Networks

ZHANG Zhao-hui

Computer Science. 2010, 37 (4): 238.

Abstract

PDF(261KB) ( 732 )

RelatedCitation | Metrics

In discrete varying-weight networks, the computational complexity of the optimal algorithms for shortest paths is far more than O(n2)and they are not suitable for dynamic traffic navigation systems. This paper presented a fast algorithm of dynamic shortest path for discrete varying weight networks. It does't select a candidate vertex as the next current one and makes a minimal sum from the current vertex to the candidate one and from it to the destination one until the current vertex is the destination. The new algorithm yields a sub-optimal solution of the shortest path, but it is more effective than the optimal algorithm and its solution is better than the dynamic solution of the shortest distance path. hhe experiments also prove that the new algorithm is applicable to dynamic traffic navigation.

Multi-objective PSO Algorithm Based on Escalating Strategy

XU Bin,YU Jing

Computer Science. 2010, 37 (4): 241.

Abstract

PDF(344KB) ( 529 )

RelatedCitation | Metrics

A multi-objective PSO algorithm based on escalating strategy was proposed. The main idea of this escalating strategy is to regenerate the whole evolutionary population with some technology, which results in a new population significantly indifferent from the old one while inheriting the evolutionary information from the history. I3y this way, the performance on global convergence can be enhanced, and premature can be avoided simultaneously. A neighborhood search procedure was imposed on some selected Pareto solutions to accelerate the evolution process for reaching a global Pareto set with well distribution. Some typical multi-objective optimization test problems were analyzed with escalation PSO and non-escalation PSO respectively to verify the effectiveness of the new algorithm. The details about how to select appropriate escalating parameters and their effect on the performance of EMPSO were also investigated to show that the EMPSO with random inertia weight factor has some advantage over than that of fixed inertia weight.

Research of Heuristic Component Functional Retrieval Based on Description Logics and Sampling Sequence Pattern Mining

YAN Wei,LIU Zhi-jing

Computer Science. 2010, 37 (4): 245.

Abstract

PDF(0KB) ( 326 )

RelatedCitation | Metrics

Recently, the Component Based Software Development has become the main direction of software development, and with the expansion of the component library, efficient retrieval is crucial for us to improve the software development. The realization of component functional retrieval was attained through semantic modeling of components,which not only made up for the lack of semantic expression, but also enhanced the retrieval efficiency. Ontology was used to describe component to create ontology library while the method of sampling sectuence pattern mining was utilined to find out frequency sequence patterns based on the historical data of retrieval, which were used to construct retrieval relationship map later. Finally, the most possible retrieval sequences were obtained to instruct users to complete component functional retrieval with high efficiency and accuracy.

Self-adaptive Crossover Particle Swarm Optimization Based on Penalty Mechanism

CHEN Jin-yin,YANG Dong-yong,LU Jin

Computer Science. 2010, 37 (4): 249.

Abstract

PDF(527KB) ( 708 )

RelatedCitation | Metrics

Particle swarm optimization (PSO) has obvious shortcoming such as local convergence, whose performance of solving constrained optimization problems needs to be improved especially in aspect of convergence speed and optimum value. In this paper, penalty mechanism based self-adaptive crossover PSO was put forward to solve the above two problems by three levels. Aiming at the local convergence problem, crossover operation was brought into PSO. Population diversity model was used to maintain population diversity to achieve global optimum solution. Penalty mechanism based self-adaptive crossover PSO was put forward which improves H strategy and simplifies P strategy. The brought algorithm has better performance without obvious algorithm complexity increasing. According to velocity variety of particles during evolution, particle moving law and the impact of parameters were analyzed. A general calculating formula was put forward to control parameters for optimizing which extends application areas for PSO.

Optimizing Support Vector Machine Parameters and Application to Fault Diagnosis

LIN Hui,WANG De-cheng

Computer Science. 2010, 37 (4): 255.

Abstract

PDF(333KB) ( 579 )

RelatedCitation | Metrics

Aiming at parameters optimization problem of classifier based on support vector machine, one kind of parameters selection algorithm was presented based on chaos genetic algorithm. The antirigonometric function Logistic map that has uniform track distribution was used to carry out chaos optimization. Therefore, it can search whole optimization interval by equal probability, overcoming disadvantage that Logistic map chaos optimization searches optimization interval edge by greater probability. Start population was produced by using chaos ergodicity. Chaos disturbance was added to chromosome that has bad fitness, in order to carry out chaos optimization. It solved premature problem and convergenee problem of genetic algorithm.This method was applied to fault classifier parameters optimization of openswitch damage in brushless do motor power converter. Experimental results assess effectiveness and feasibility of the proposed approach.

Design and Implement of UAVs Testbed Based on 3D Environment

ZHOU Wei,WEI Rui-xuan

Computer Science. 2010, 37 (4): 258.

Abstract

PDF(367KB) ( 812 )

RelatedCitation | Metrics

This paper described recent research on the design and implement of Muti Agent testbed for UAVs cooperative control experiment in order to improve the reality and universality based on 3D environment. First, the testbed architecture was designed by HLA standard. Secondly, the specification was implemented from three step including software system architecture design, runnable flow design, interface design, the system was implemented using VC++tool, LOD algorithm, MAPX and UAV model. The system was proved reality and universality by simulation.

Binary Image Representation Method Using Improved TNAM

FANG Shao-bong,ZHENG Yun-ping,CHEN Chuan-bo

Computer Science. 2010, 37 (4): 261.

Abstract

PDF(346KB) ( 689 )

RelatedCitation | Metrics

A binary image representation method using the improved TNAM (Triangle Non-symmetry and Anti-pateking pattern representation Model) , which is called the ITNAM representation method, was proposed. Also, a concrete algorithm of the ITNAM for binary images was presented and the storage structure, the total data amount, and the time and space complexities of the proposed algorithm were analyzed in detail. By taking some idiomatic standard images,such as‘Lena'，`Baboon' and‘Peppers',in the field of image processing as some typical test objects, and comparing the algorithm of the IhNAM with those of the classic linear quadtree (LQT) and the latest TNAM,the theoretical and experimental results show that the former is obviously superior to the latters with respect to the numbers of subpatterns or nodes and the data storage, and therefore it is a better method to represent binary images.

ASM and Color Gabor Features for Facial Feature Extraction

ZHU Jie,TANG Zhen-min

Computer Science. 2010, 37 (4): 265.

Abstract

PDF(350KB) ( 584 )

RelatedCitation | Metrics

To improve active shape Modcl(ASM) accuracy in facial feature points location in facial images, an improved ASM based algorithm was proposed. First, the irises were localized and utilized to initialize the shape model. Second,global face features with salient features were employed to constrain the movement of feature points; at last, in order to improve ASM, Color Gabor features was used to extract edges and corner points for feature, so we could get the key facial feature points quickly and accurately. Experimental results show that our algorithm performs significantly better than the traditional ASM

Real-time Electronic Digital Image Stabilization System Based on Multi-core Computation

ZHOU Yu-bin,SHI Yong-gang

Computer Science. 2010, 37 (4): 269.

Abstract

PDF(540KB) ( 685 )

RelatedCitation | Metrics

The multi-core computing has become the new trend of computer industry. Electronic Digital Image Stabilization (EDIS) system involves both large amount of data and intensive computation. The real-time implementation of EDIS system entails rapid data processe with high parallelism and repeated computation,which are suitable for data parallel processe of image sectuences. The algorithm solution was proposed to deal with high speed image stabilization to meet the requirements of real-time systems. According to the data processe characteristics of EDIS, efficient C++ with inline SIMI)(Single Instruction Multiple Data) and multi-thread programming were written and compiled on dual-core PC for real parallel running threads to achieve real-time performance in conventional PC. After the motion vectors acquired from fast search strategy, adaptive multi-local motion vectors avoided the moving objects interference. And the Kalman filter removed high-frequency image fitter motion while smooth global movements were retained for scanning and vehicle moving. The test results show that it is possible to implement a high efficient and robust real-time stabilization system on conventional PC with Multi-core at last.

Denoising Algorithm of Proportional Shrinkage with Enhancement Based on the MAP Rule in Wavelet Domain for Infrared Image

LIU Gang,LIANG Xiao-geng,LUO Xu-tao

Computer Science. 2010, 37 (4): 274.

Abstract

PDF(342KB) ( 568 )

RelatedCitation | Metrics

In order to solve the fault of weakening the detail and edge of image while denoising in wavelet domain, this paper presented an adaptive denoising algorithm with detail enhancing and applied it to infrared image. On the basis of the assumption that the prior distribution of the original image coefficients and the noise coefficients were both Gaussian,this method firstly made use of the rule of Maximum a Posteriori to compute the shrinkable factor for wavelet coefficients, then revised it by taking decomposable level and directional energy into account. Finally, a denoising and enhancing image could be obtained when the wavelet coefficients which were shrunk by the revised shrinkable factor experienced the reverse transform. The experimental results show that the method given by this paper, compared with the direct proportional shrinkage,can enhance image's detail and improve image's contrast and get better visual effect though it has a little loss of Peak Signal Noise Ratio. The idea of coefficients' enhancement in wavelet domain proposed by this paper can apply to other proportional shrinkable algorithms.

Improved Interpolation Algorithm for Demosacing Bayer Pattern Images

YIN Yong,HU Lei

Computer Science. 2010, 37 (4): 278.

Abstract

PDF(265KB) ( 893 )

RelatedCitation | Metrics

Image sensors can only gain the Bayer pattern pictures which arc filtered by CFA. In order to get the full colour information, the other two missing colour samples must be interpolated. An algorithm based on Hamilton-Adam interpolation and trigonometric function rule was presented, which can satisfy the demand of real-time system, and can be implemented by hardware easily. Then statics and simulation pictures produced by matlab were given to estimate this algorithm,and the difficulty of the implement for the algorithm on hardware was discussed at the end. The results of experiments proved that the algorithm can restore the edges of the pictures better, and can provide satisfying speed when implied on hardware.

Face Detection Based on Adaptive Tracking Window Scale

LEI Zhen,WANG Qing-hai,WU Ling-da,XUE Ting-mei

Computer Science. 2010, 37 (4): 281.

Abstract

PDF(325KB) ( 554 )

RelatedCitation | Metrics

As an effective iterative algorithm, Mean-shift is widely used in object tracking, image smoothing and other computer vision areas. In view of the fact that the standard Mean-shift tracking algorithm is lack of scale adaptation mechanism and Camshift algorithm needs choosing face local sample manually to track accurately before detecting,an adaptive tracking window scale algorithm was proposed for detecting face in video. I}he algorithm adopts illumination compensation and skin color segmentation to rectify the size and location of tacking window. The results of experiments show that the proposed algorithm not only has good real-time capability, but also reduces the location error of the traditional algorithm and achieves more accurate face detection in video.

Analysis of Digital Rubbings Key Technology

ZHU Xiao-li,WU Zhong-fu,LI Ying

Computer Science. 2010, 37 (4): 285.

Abstract

PDF(431KB) ( 818 )

RelatedCitation | Metrics

Digital rubbing is a new approach to inherit rubbings art.The key of digital rubbings is how to solve the uneven illumination of stone images and black background,to generate ideal images that can not only full reflect target images and details,but also have the basic characteristics of the rubbings in the process of making digital rubbings.Through specific examples,this paper analyzed the basis for confirming the key technology of digital rubbings.The experiment results show that these key technologies have solved the above problems properly and produced ideal digital rubbings.

Real-time Modeling and Rendering on the Realistic of Snow Scenes

TANG Yong,ZHANG Qian

Computer Science. 2010, 37 (4): 289.

Abstract

PDF(372KB) ( 876 )

RelatedCitation | Metrics

Snow meticulous rendering is one of the challenging topics in the field of computer graphics. A method for real time realistic rendering of snow sconce was proposed .First of all, the track of attenuation function as snowfall was applied to snow scenes model based on visual snowfall, which was presented to ensure authenticity at the same time to improve its efficiency. Then real-time accumulating-melting snow model was presented based on temperature change and surface types, by calculating the lossless heat transfer equation, interactive dynamic simulation of the snow scene could be improved. Finally the GPU graphics hardware acceleration technology was used to enhance the efficiency of the whole calculating process. Rendering results arc realistic.

Face Detection System Design Based on Two Classifiers

ZHANG Yong,XUE Zhi-mao

Computer Science. 2010, 37 (4): 293.

Abstract

PDF(307KB) ( 573 )

RelatedCitation | Metrics

The face recognition technology has broad application prospects,but right now there are some failings in many implementations.Based on the analysis of OPENCV with the SVW classifier,this paper elaborated a strategy based on two-stage classifier of face detection principle and the implementation process.first,analysed two classifier construction,introduced a rectangular face image feature vector,used the image of the rectangular features as the basis for classification,and then discussed the system design and implementation,including the gray-scale transformation process,the process of histogram equalization,image smoothing serialization process as well as the relization of the pyramid.This test model can speed up the processing speed and has higher efficiency.

Studies on Dynamic Adjustment Mechanism of Real-time Video Streaming on Surveillance System

BA Hong-tao,CHEN Yao-wu,ZHOU Rong

Computer Science. 2010, 37 (4): 296.

Abstract

PDF(251KB) ( 647 )

RelatedCitation | Metrics

In order to design a real-time and fluent video surveillance system, this paper presented a new R TCP-based feedback congestion control algorithms to make up for the deficiencies of current way using the congestion control algorithms. In the algorithm, utilizing the key parameters mate the circle time and packet loss rate,forecast the in the feedback TCP packet received from the client, we can network bandwidth to adjust the encoding bit rate of server-side video stream.Experiments show that the algorithm can adjust the video stream encoding bit rate dynamically according to network conditions so that the client can get the smooth real-time video stream.