Current Issue

• 2018, No. 11　Published:25 November 2018 Last issue    Next issue
  Columns in this issue:

• Remote audit of big data cloud storage based on

   finite field algebraic signature partition table

  QIAN Zheng1,2,XIA Hongxia2

  2018, 40(11): 1907-1914. doi:

  Abstract ( 107 )   PDF (777KB) ( 173 )   　　 Review attachment

  In order to improve the audit efficiency of big data storage, we propose a big data storage audit method based on the remote data checking (RDC) of the finite field algebraic signature partition table. Firstly, after obtaining the algebraic signature of outsourced files, we use the arithmetic operation of the underlying field to complete the remote detection of data integrity in the cloud storage. The proposed data audit method has relatively low computation and communication costs for both the client and the cloud server. Secondly, we design a divide and conquer table (D&CTs) as a new data structure to effectively support dynamic data operations such as insert, add, delete, and modify operations. The D&CTs method can enable the proposed RDC scheme to be applied to cloud storage process analysis with varied size of files. Finally, simulation experiments are carried out to verify the effectiveness of the proposed method in big data cloud storage process.

   

  Algorithms for task joint scheduling and computation

  offloading in mobile cloud computing

   

  LUO Yuchun,WU Jigang,SHI Wenjun,HE Zinan

  2018, 40(11): 1915-1924. doi:

  Abstract ( 143 )   PDF (1008KB) ( 180 )   　　 Review attachment

  With the development of the Internet, many applications have a growing demand for computing power and resources. However, mobile devices have limited resources, such as battery life, network bandwidth, storage capacity, and processor performance. Cloud offloading is a main solution to supporting computationally demanding applications on these resource constrained devices. We propose a fast and efficient heuristic algorithm for the scheduling and offloading problems of the application tasks in the wireless network. The heuristic algorithm initially moves the tasks which can be offloaded to the cloud, then successively calculates the energy saving of each offloaded task running on the mobile terminal, and sequentially moves the tasks with the highest energy saving to the mobile device. The saved energy is updated in  each iteration in order to cater for the task concurrence. In addition, we also construct a simulated annealing algorithm,  which uses the solution generated by the heuristic algorithm as the initial solution, to further optimize the solution obtained by the heuristic algorithm, and depict in detail the encoding method, objective function, neighborhood solution,  temperature parameters, and algorithm termination rules. Experimental results show that in comparison to the three algorithms based on nonoffloading, full offloading, and random offloading respectively, the solution generated by the heuristic solutions is better and efficient.

   

  A reliability aware task scheduling

   algorithm for cloud computing

  QI Ping1,2,WANG Fucheng1,WANG Biqing1,LIANG Changyong2

  2018, 40(11): 1925-1935. doi:

  Abstract ( 169 )   PDF (830KB) ( 167 )   　　 Review attachment

  Parallel tasks in the cloud computing environment are vulnerable to resource failure and hence cannot be completed, and dynamically providing cloud resources has low reliability. Aiming at this issue, firstly, we introduce a failure recovery mechanism. Because the failure regularity of resources changes dynamically under the condition of failure recoverability, the twoparameter Weibull distribution is used to describe the local characteristics of resource nodes and the failure regularity of communication links in different time periods. Then, based on the analysis of various interactions between parallel tasks, we propose a resource reliability evaluation model based on variableparameter failure regularity. Finally, the model is incorporated into the particle swarm optimization algorithm to obtain the reliabilityaware and adaptive inertia weight PSO resource scheduling algorithm (RPSO), so that the reliability of the alternative resources is fully considered when calculating the fitness. Simulation results show that when appropriate failure recovery parameters are selected, the proposed RPSO algorithm can increase the reliability of cloud services and only add a small amount of additional failure recovery overhead.

   

  An improved adaptive random testing method

  in high dimensional input domains

   

  ZHAN Xuzheng

  2018, 40(11): 1936-1943. doi:

  Abstract ( 116 )   PDF (491KB) ( 142 )   　　 Review attachment

  Adaptive random testing (ART) ensures that test cases are more evenly distributed in the input domain, and thus achieves significantly stronger failure detection capability than the basic random testing.  Among the existing ART methods, the fixedsizecandidateset ART (FSCSART) exhibits better failure detection capability and has extensive applications. However, its failure detection effectiveness deteriorates significantly with the increase of input domain dimensions. To solve this high dimension problem, two types of distances are taken into account while choosing a test case from the candidate set: one is the distance from each candidate point to the already executed test cases; the other is the distance from individual candidate point to the center point. The comprehensive consideration of distances can reduce the priority of the candidate points at the edge of the input domain and overcome the disadvantage of the FSCSART. Experimental results show that the improved algorithm achieves a stronger failure detection capability in high dimensional input domains.

   

  An analog circuit fault diagnosis method

  based on DCQGA-SMKL-SVM

  YAN Xuelong，GONG Liuqing，WANG Binbin

  2018, 40(11): 1944-1950. doi:

  Abstract ( 126 )   PDF (629KB) ( 158 )   　　 Review attachment

  We present an analog circuit fault diagnosis approach by using the double chain quantum genetic algorithm to optimize the simple multi-kernel learning support vector machine (DCQGA-SMKL-SVM). Firstly, we extract the time domain response signals of the test circuit, and utilize the Harr wavelet to process and normalize response signals to obtain characteristic parameters. Then, the parameters of the SMKL-SVM are optimized by the DCQGA and an analog circuit fault diagnosis model based on the DCQGA-SMKL-SVM method is constructed for analog circuit fault diagnosis. Tests on the bi-quadratic filter circuit and quad-amplifier second-order high-pass filter circuit show that compared with the DCQGA-SVM method, the proposed approach has higher fault diagnosis accuracy.

   

  A new method for generating signals with

  variable frequency and adjustable waveform

   

  HU Anfeng,WANG Suzhen,SHEN Zhongjie，WANG Wei,ZHANG Hao

  2018, 40(11): 1951-1956. doi:

  Abstract ( 85 )   PDF (2402KB) ( 138 )   　　 Review attachment

  Direct digital frequency synthesis (DDS) technology is a new type of frequency synthesis technology. It has a higher frequency precision, which can not only achieve frequency switching quickly but also ensure the continuity of the phase when the frequency is changing. However, the output waveform and frequency range of special DDS chips are usually fixed. We extend the structure of DDS circuits by adding some modules such as data distributor, ROM for storing different waveform data, and peripheral control circuit module based on the research on special DDS chip circuits. Thus variablefrequency multimode signals are generated. The whole variablefrequency signal system is integrated on the large scale programmable FPGA chip to achieve waveform programming and frequency programming. Meanwhile, it can not only realize waveform selection such as sine wave, triangular wave, sawtooth wave, square wave, etc., but also change the frequency of each waveform. The system integrates many modules such as the PLL frequency multiplier, frequency divider, data selector, data distributor, frequency word input module, DDS signal generator, and key module on the programmable FPGA chip, which greatly improves the integration and reliability of multimode variablefrequency signal circuits. Due to the programmability of the FPGA, the system parameters can be adjusted through field programming, which increases the flexibility of circuit adaptation.

   

  
   

   

  A DDoS attack security situation assessment model

  based on improved fuzzy C-means clustering

  ZHANG Ruizhi1,TANG Xiangyan1,CHENG Jieren1,2

  2018, 40(11): 1957-1966. doi:

  Abstract ( 89 )   PDF (792KB) ( 148 )   　　

  Traditional network situation assessment methods cannot effectively evaluate the distributed denial of service (DDoS) attack security situation in the new network environment. We propose a DDoS attack security situation assessment model based on improved fuzzy C-means (FCM) clustering. This model generates a fusion feature gained from network flow IP address changes of old and new users and unilateral and bilateral network flow, and calculates the risk indexes of each network node on the basis of the fusion feature. The security situation information of the whole network can be obtained by fusing the risk indexes of all the nodes in the network, which is then classified into five security levels by the improved FCM. The DDoS attack security situation of the whole network therefore can be quantitatively evaluated by the proposed model. Experiments on real DDoS data show that the proposed model can assess the DDoS attack security situation reasonably and effectively, and it is more flexible and accurate than existing methods.

   

  A fast recovery algorithm for survivable

  spanning trees in networks

   

  ZHENG Lulu,WU Jigang,CHEN Hui

  2018, 40(11): 1967-1973. doi:

  Abstract ( 90 )   PDF (562KB) ( 130 )   　　

  How to cope with the failures of network links is one of the most challenging problems. Existing research usually adopts a survivable spanning connection (SSC) with two spanning trees to avoid link failures. Due to the rapid growth of network data transmission rate, the spanning trees in SSC all become invalid when the shared links of a pair of two spanning tree fail. We propose a fast recovery algorithm for survivable spanning trees in case of shared link failures in SSC. The algorithm searches the replaceable link set of invalid links and adds the links with least probability of failure into the spanning trees in the original SSC, thus generating a new SSC. Experimental results show that the proposed algorithm can keep the close-to-optimal survivability of the network while significantly reducing recovery time and time complexity. The recovery time is optimized by up to 34.42% when the number of network nodes changes from 10 to 100, and the error in survivability does not exceed 1%.

  An improved stable AODV protocol

  scheme based on fuzzy neural networks

  HUANG Baohua1,MO Jiawei2,L Qi1

  2018, 40(11): 1974-1982. doi:

  Abstract ( 104 )   PDF (757KB) ( 136 )   　　 Review attachment

  One of the important characteristics of vehicular ad hoc networks is the high mobility of nodes. It is significant to select stable links for transmitting data in routing protocols to solve the problem of frequent link breakage caused by free movement of nodes. To solve this problem, we propose an improved ondemand distance vector routing protocol (AODV) scheme with link stability, namely GFAODV (AODV with GASAFNN). The algorithm uses fuzzy neural networks to calculate the node information in the initialization and selection stages to get node stability. Then we evaluate the link quality, and select the stable link which has fewest hops based on link stability and hop count. In the routing maintenance phase, the parameters of the fuzzy neural network are optimized in real time according to the actual environment by using the genetic simulated annealing algorithm, which guarantees the consistency between calculated node stability and the actual situation. Experiments show that the GFAODV routing protocol outperforms the AODV in terms of average delay, packet delivery rate and routing overhead.

   

   

  Key node identification methods based on road traffic networks

  YAN Kai1,2,LI Ling1,2,QIN Yongbin1,2

  2018, 40(11): 1983-1990. doi:

  Abstract ( 123 )   PDF (546KB) ( 314 )   　　 Review attachment

  In real world, a large number of complex systems can be characterized by abstract nodes and networks of connected edges. As an important part of the urban transportation system, the road transportation network is a typical complex system, which is closely related to people's lives. The key nodes identification problem in road traffic networks is a classic problem in complex networks. The traditional degreecentric algorithm and PageRank algorithm are extensively in use for the identification of key nodes of complex networks. Considering the particularity and correlation of key nodes in the road traffic network, we introduce the idea of the greedy algorithm into the degreecentric algorithm, and propose a key nodes identification method.  We also introduce the idea of the greedy algorithm into the PageRank algorithm, and propose a key node identification method. The results of identifying key nodes in the road traffic network by the proposed methods are more reasonable, which means they have important application value in the fields of traffic road maintenance, planning and design, and prevention of criminals' escape. Experiments on public datasets verify the effectiveness of the two proposed algorithms in comparison with classic key node identification methods.

  A link quality estimation method

  based on closeness grades

  ZHANG Hejie,MA Weihua

  2018, 40(11): 1991-1999. doi:

  Abstract ( 125 )   PDF (897KB) ( 196 )   　　

  In wireless sensor networks (WSNs), volatility of the link affects the accuracy of data transmission of the upper routing protocols. To improve the efficiency, link quality evaluation is used to avoid choosing poor links and increase the efficiency of routes. Link hierarchy grading in current link quality estimation study is subjective without unity. Aiming at this problem, we use the entropy method to calculate the weight of evaluation parameters to eliminate the interference of subjective factors in the calculation. Since the link quality is affected by multiple feature attributes, we then determine link quality grades by the closeness analysis method. According to the grades, we propose a link quality evaluation method based on closeness grades, which uses the dispersion degree of classes to establish a binary decision tree for classifying link quality. We also build a four level binary decision tree estimation model of link quality based on the support vector machine (SVM). Besides, we utilize a hybrid algorithm to optimize the parameters of the kernel function. Experimental results indicate that the improved algorithm can increase estimation accuracy  with less training time. Comparison in multinetwork scenarios shows that the proposed model outperforms the conventional link quality estimation model based on LQI and the estimation model based on BP neural network. It can accurately assess current link quality with a small number of probe packets, thus reducing energy consumption with good adaptive capacity to the environment.

  Survey on software vulnerability analysis

  based on machine learning

  KUANG Xiaohui1,LIU Qiang1,2,LI Xiang1,NIE Yuanping1

  2018, 40(11): 2000-2007. doi:

  Abstract ( 249 )   PDF (544KB) ( 469 )   　　 Review attachment

  As increasing reporting and disclosure of vulnerability code samples and extensive applications of machine learning methods, software vulnerability analysis methods based on machine learning have become a hot research direction in information security. After analysis of existing research work, we propose a software vulnerability analysis framework based on machine learning. We then review and classify existing machine learning based vulnerability methods, and conduct comparative analysis. We briefly analyze the challenges for machine learning based software vulnerability analysis methods, and discuss future research trends.

   

  Symbolic CTL model checking based on possibility measure

  LEI Lihui，GUO Yue，ZHANG Yanbo

  2018, 40(11): 2008-2014. doi:

  Abstract ( 104 )   PDF (546KB) ( 158 )   　　 Review attachment

  With the increasing complexity of systems, it is urgent to deal with the uncertain information in the systems. Besides, the state explosion problem is getting more and more serious. Existing model checking techniques are no longer suitable for the verification of such complex systems. We study symbolic CTL model checking based on possibility measure. Firstly, multiterminal binary decision trees (MTBDDs) and Boolean formula are respectively used to describe system models and properties to be verified. Secondly, the system models are normalized and simplified. Finally, system verification is completed by fixed point calculation. Our work is an integration of symbolic model checking and CTL model checking based on possibility measure, which can not only handle uncertain information in systems, but also maintain symbolic model checking's advantage of low demand for computation time and storage space. It is thus significant for the verification of complex systems.

   

  Illumination robust face recognition based on

  random projection and weighted residual of

  sparse representation based classification

  LI Yan，ZHANG Yue

  2018, 40(11): 2015-2022. doi:

  Abstract ( 81 )   PDF (654KB) ( 170 )   　　

  To solve the problem of illumination variation in face recognition, we improve the traditional sparse representation classification (SRC) by using random projection, and propose an illumination robust face recognition method based on random projection and weighted residual of sparse representation based classification (RPWRSRC). By normalizing the illumination of the face image, the bad illumination on the face image is eliminated as much as possible. Then we introduce random projection into the proposed method. The face images with illumination correction are projected to multiple random spaces to enrich the illumination invariant features of the samples and to reduce the influence of illumination changes on face recognition. On this basis, we utilize the single residual classification to improve  the traditional sparse representation classification method. By multiple random projections and sparse representations, we can obtain multiple sample features and reconstructed residuals. With the energy of sample features, we determine the weights for fusion of respective reconstructed residual to get a more stable and reliable weighted residual instead of single residual in traditional SRC. Experiment results on Yale B and CMU PIE face databases show that, the proposed RPWRSRC has strong robustness for face recognition under different illumination conditions. Compared with the traditional SRC, its average recognition rates of the two experiments on Yale B face database are increased by 25.76% and 46.39% respectively, and the average recognition rate on CMU PIE is increased by about 10%.

  An optimal utility selection algorithm for

  crowdsourcing  photos based on metadata

   

  SONG Jieqiong,ZHAO Ming

  2018, 40(11): 2023-2032. doi:

  Abstract ( 96 )   PDF (1326KB) ( 137 )   　　 Review attachment

  With the rapid development of the Internet, crowdsourced photos generated by mobile terminal devices can be used in many critical application scenarios, such as post earthquake recovery and emergency management. However, such applications often have resource constraints such as bandwidth, terminal device storage and processing capability, which limit the number of crowdsourced photos. Thus, it is a big challenge to use the limited resources to best reproduce the target from crowdsourced photos. In this paper, we collect and process various geographical and geometrical data of the photos to form a metaarray of photos, and propose an optimal utility selection algorithm of crowdsourced photos to achieve the best reproduction of the target under the condition of limited resources. The input of the algorithm is metadata rather than pixels, so it can efficiently analyze crowdsourcing in resourceconstrained applications. The utility of photos is used to measure the coverage degree of the target area, and an effective photo utility calculation method is proposed. Finally, we design a simulation experiment, and the results prove the effectiveness and superiority of the algorithm.

   

  An optical compensation based quick

  dehazing algorithm using dark channel prior

  YANG Yan,ZHANG Baoshan,ZHOU Jie,CHEN Gaoke

  2018, 40(11): 2033-2039. doi:

  Abstract ( 147 )   PDF (1008KB) ( 170 )   　　

  Aiming at the problems of incomplete dehazing, color cast effect in large sky regions and low efficiency of the single image dehazing algorithm using the dark channel prior, we propose a new optical compensation based quick dehazing algorithm using dark channel prior. Firstly, we introduce the twoorder Butterworth high pass filter into the homomorphic filtering function, and enhance the minimum color components in the frequency domain. Meanwhile, smoothing the illumination in the minimum color component compensates for the degradation of image quality caused by insufficient illumination in the local area. Secondly, we smooth the above image by bilateral filtering, which can make illumination in minimum color components more natural. Finally, the processed minimum color components are regarded as the guide image to refine the initial transmittance. Experimental results show that compared with the Tarel algorithm and median filter algorithm, the restored images processed by the proposed algorithm have better visual effect. Compared with the guided filter algorithm, the recovered hazefree images processed by the proposed algorithm has  better dehazing effect and more accurate color restoration in the sky region, and faster computational speed as well.

   

  An SVM classifier based on chaotic gray

  wolf optimization algorithm

  WANG Zhihua,LUO Qi,LIU Shaoting

  2018, 40(11): 2040-2046. doi:

  Abstract ( 180 )   PDF (1284KB) ( 239 )   　　

  The support vector machine (SVM) is a small computational data set established under the classification problem, which can achieve nonlinear highlatitude classification with good scalability. However, during the training process of traditional SVM, the results of SVM computation are closely related to parameter selection, and the parameter selection algorithms currently in use have a number of defects. Aiming at above problems, we introduce the gray wolf algorithm (GWO) into the chaotic sequence, change the initial distribution of wolves, and propose a chaotic gray wolf optimization algorithm (CGWO), which can improve the uniformity of wolf distribution and the ergodicity of wolf searching, thus greatly enhancing the computing speed and accuracy of the GWO algorithm, and ultimately achieve better SVM optimization. Comparative experiments on the open source data provided by Mirjalili mixed with the original data as the test set of the vector machine shows that the CGWO algorithm has obvious performance improvement, and it outperforms the GWO algorithm,  artificial bee colony, gravitational search algorithm and SVM optimized by traditional optimization algorithms, with higher computation accuracy  lower error and less time.

   

   

  Failure rate prediction based on

  grey multiple linear regression model

  GUO Lijin1,HE Xishuo1,2,3,XU Xinxi2,SHI Meisheng2,WANG Jihu3

  2018, 40(11): 2047-2053. doi:

  Abstract ( 131 )   PDF (535KB) ( 221 )   　　 Review attachment

  We propose a grey multiple linear regression model to predict the failure rate of the oxygen system during the using year of oxygen equipment. Firstly, we find out the GM (1,1) model of the failure rate of the oxygen system equipment. Secondly, we calculate the relationship model of the oxygen system failure rate, oxygen equipment failure rate and the years of equipment in use, and plug the GM (1,1) model of oxygen equipment failure rate into the relational model. Finally, we calculate undetermined parameters by using the least square method. We analyze the failure rate prediction of the oxygen system, and the results show that the grey multiple linear regression model is superior to both the individual GM model and linear regression model in terms of prediction accuracy of failure rate. Moreover, the historical data in use does not require a typical distribution. The prediction results of the model can provide a decisionmaking basis for oxygen system maintenance work.

   

  An improved differential evolution algorithm based

  on multi-generation population evolution information

  #br#  

  SONG Qiang1,LIU Yaping2,LIU Zhenlan1

  2018, 40(11): 2054-2059. doi:

  Abstract ( 102 )   PDF (423KB) ( 179 )   　　

  The differential evolution algorithm is one of the global numerical optimized algorithms with excellent performance, and it has been widely applied in artificial intelligence, signal processing and so on. However, the current research takes into account the population distribution information of  generation and neglects the distribution information accumulated by the multigeneration cumulative population in the evolution process, thus the distribution information is not fully utilized. Inspired by  covariance matrix adaption evolutionary strategy(CMAES), we propose a new method that can make full use of the accumulated population distribution information in the evolution process. As the CMA tends to premature converge and falls into local optimum, the proposed method improve the mutation operation and cross operation to balance global and local search capability. Firstly, we sort the vectors according to their fitness value, and calculate the probability of individual vector participating in mutation operation based on the probability model improved by cosine function. To improve the global search ability, the proposed method selects the base vectors and the end vectors of difference vectors by their probability value in descending order, and the initial vectors of difference vector are selected in ascending order. Then, we establish the new coordinate system by eigenvectors which are generated from the decomposition of the covariance matrix. Executing crossover operation in the new coordinate system makes the trial vector closer to the global optimum than the traditional way. Experimental results show that, the proposed algorithm outperforms the existing improved algorithms on test function IEEE CEC2014.

  Review:Detection of product review spam

  ZHANG Sheng,WU Xing,ZOU Dongsheng

  2018, 40(11): 2060-2066. doi:

  Abstract ( 93 )   PDF (489KB) ( 155 )   　　 Review attachment

  E-commerce website's product reviews have a guiding effect on the user's purchase intention. However, spammers may create fake reviews to artificially promote or demote target products. Techniques for detecting review spam therefore have become an urgent need. We first divide the related researches into two categories: internal review (text features) -centric and external reviews (behavior features) -centric. Then, their research progresses on feature selection and machine learning methods are summarized. The common data sets in product review spam detection are summarized. Finally, we point out some potential research directions based on the review.