Showing papers in "Application Research of Computers in 2013"

Bat algorithm with global convergence for solving large-scale optimization problem

Abstract: To solve large-scale optimization problems(OP),this paper constructed a bat algorithm with global convergence.In the algorithm,each bat was just an alternative solution of OP,used the principle of orthogonal Latin squares to construct initial positions of bat group so as to cover search space with balance dispersion and neat comparability,used the following behavior,autonomous behavior,averting-danger behavior and conformability behavior of bats to construct space positions trasnsfering strategies;used the loudness and rate of plus emission to ensure the bat groups keep either to stay unchanged or to transfer toward better positions,but never to transfer worse positions.During evoluation process,a bat's transferring from one position to another realized the search for the optimium solution.The result shows that the reducible stochastic matrix stability theorem can be applied to prove the global convergence of the algorithm.The case studys show that the algorithm has advantages of good suitability for different types of OPs and high convergence speed when applied to solve large-scale optimization problems.

Wolf colony search algorithm based on leader strategy

Abstract: This paper presented a wolf colony search algorithm based on the leader's strategy,which was based on the characteristics of the wolves prey behaviorThe algorithm ideas from the following processWolves' individuals competed with each otherSo,the strongest wolf was selected as the leader of the wolves,the wolves hunted prey under the leadership of the leader,so that they could be more effective to capture preyUltimately wolves could find the global optimum solution by constantly search and capture prey under the leadership of the leader wolfIn order to prove the algorithm's validity,it used test standard functions to make comparison with other algorithmsThe results show that the algorithm has better performance in aspects of convergence speed and accuracy,and it is difficult to trap-in local minimum

Fruit fly optimization algorithm with adaptive mutation

Abstract: In order to overcome the problems of low convergence precision and easily relapsing into local extremum in basic fruit fly optimization algorithm(FOA),this paper presented an adaptive mutation fruit fly optimization algorithm(FOAAM).During the evolution,in the condition of basic FOA's trapping in local extremum judging from the population's fitness variance and the current optimal,first,it generated M current optimal replicates.Then,it disturbed replicates by a certain probability P Gauss mutation operator.Finally,it optimized mutated replicates again to jump out of local extremum and continue to optimize.Experimental results show that the new algorithm has the advantages of better global searching ability,speeder convergence and more precise convergence.

Predict algorithm of micro-blog retweet scale based on SVM

Abstract: Based on the analysis of the factors that affect retweet behavior,this paper proposed a predict SVM algorithm with five features:user influence,user activity,interest similarity,the importance of micro-blog content and users closeness.Furthermore,it proposed the predict algorithm of retweet scale on the basis of SVM,also,gave a method to evaluate the predict accuracy.The experiment with Sina micro-blog data shows a good result that the predict accuracy is up to 86.63%.

Survey of feature representations and similarity measurements in time series data mining

Abstract: This paper respectively analyzed the function and meaning of feature representations and similarity measurements for time seriesIt also summarized the existed methods and analyzed the merits and demeritsMeanwhile,by discussing the noteworthy problems,it provided the further research direction of feature representations and similarity measurements for time series

Survey on software defined network research

Abstract: Due to the support of enormous network protocols,IP network device results in high complexity.Not only does it limit the IP network technology development,will not be able to meet the current application trend of cloud computing,big data and server virtualization.Software defined networking(SDN) as a latest network norm making network device programmable with the Abstraction redefinition of control plane is expected to solve above issues.This paper introduced the background,characteristics and development status of OpenFlow,it analyzed OpenFlow-based SDN architecture and key technologies.Furthermore,it discussed different SDN research directions,such as the automation of network management,the unified control of optical transmission and IP bearing,smooth switching in wireless network,network virtualization and QoS assurance.

Images matching algorithm based on SURF and fast approximate nearest neighbor search

Abstract: This paper proposed an images matching algorithm based on SURF and fast approximate nearest neighbor search for that nearest neighbor matching of high-dimensional feature vector was low.First,this algorithm used Fast-Hessian detection to find features,and generated feature vector of SURF descriptors.Then using bidirectional approximate nearest neighbor matching algorithm to match,finally adopted PROSAC algorithm to exclude mistake matching points.Experiments show that the algorithm not only improves the matching correct rate of SURF algorithm,and ensure the real-time nature.

Research of 3D reconstruction and filtering algorithm based on depth information of Kinect

Abstract: This paper analyzed and improved 3D reconstruction algorithm using the depth information from Kinect.To reduce noise,it proposed an improved bilateral filtering algorithm based on the signal structure.This new algorithm used a two-valued function to compute the weights of the filter,because the range of depth image data was already known.It also combined the RGB values and depth information of surrounding pixels to complement some missing depth information.The results show that the proposed algorithm has much better performance and efficiency,namely 6 times as fast as the original algorithm.

Coverage optimization of wireless sensor networks based on artificial fish swarm algorithm

Abstract: Focusing on the problems of increasing network cost and short life cycle caused by serious redundant network nodes,this paper proposed a coverage optimization method based on artificial fish swarm algorithm.Firstly,it took the nodes' utilization and the efficiency of network coverage as the optimization goal to establish mathematical model.And then it used the artificial fish swarm algorithm to solve the model,and got the optimal coverage scheme for WSN.The simulation results show that the proposed method can not only improve the coverage rate of WSN,but also reduce redundancy and the cost of network effectively,so that the network life time has been extended.

Study on network traffic forecast model of SVR optimized by GAFSA

Abstract: There are some problems,such as low precision,on existed network traffic forecast model.In accordance with these problems,this paper proposed the network traffic forecast model of support vector regression(SVR)algorithm optimized by global artificial fish swarm algorithm(GAFSA).GAFSA constituted an improvement of artificial fish swarm algorithm,which was a swarm intelligence optimization algorithm with a significantly effect of optimization.The optimum training parameters could be calculated with optimizing by chosen parameters,which would make the forecast more accurate.With the optimum training parameters searched by GAFSA algorithm,a model of network traffic forecast,which greatly solved problems of great errors in SVR improved by others intelligent algorithms,could be built with the forecast result approaching stability and the increased forecast precision.The simulation shows that,compared with other models,the forecast results of GAFSA-SVR network traffic forecast model is more stable with the precision improves to more than 89%,which plays an important role on instructing network control behavior and analyzing security situation.

Survey on algorithms of low-rank matrix recovery

Abstract: This paper collectively referred robust principal component analysis,matrix completion and low-rank representation to as low-rank matrix recovery,and made a brief survey on the existing algorithms of low-rank matrix recovery.Firstly,it discussed various optimization models and the corresponding iterative algorithms for robust principal component analysis.Next,it analyzed the matrix completion problem and proposed the inexact augmented Lagrange multipliers algorithm to solve the problem.In addition,it introduced the optimization models for the low-rank representation problem and presented the iterative algorithm.Finally,this paper discussed several problems which need further research.

Artificial glowworm swarm optimization algorithm for 0-1 knapsack problem

Abstract: According to the principle of swarm intelligence,this paper proposed a new optimization algorithm based on the ideas of glowworms:the glowworm swarm optimization(GSO) algorithm to solve the 0-1 knapsack problem.Through the numerical simulations,it compared with that of artificial bee colony algorithm,ant colony optimization algorithm and particle swarm optimization.And it obtains the satisfactory results,which show the validity and effectiveness of the algorithm,expands the applications of GSO.

Distribution location routing optimization problem of food cold chain with time window in time varying network

Abstract: In order to solve the problem of food cold chain logistics distribution system optimization problem,for perishable goods characteristics,combined with the distribution network time-varying characteristics to analyse travel time,this paper designed satisfaction degree function according to service time windows and established the simulation model under time-dependent.It designed the two-phase solution of preoptimization phase and real-time optimization phase,by using the decomposition method,it decomposed the problem,designed the minimum envelope clustering analysis method and tabu search algorithm to solve the problem.Simulation results show the effectiveness of the model and algorithm of practical value.

Survey on motion capture technique and its applications

Abstract: Motion capture technique has been widely used in many application fields for the capable of measuring,tracking and recording the motion trajectory of objects in 3D spaceAfter introducing the development process of motion capture technique,it summed up 5 types of commonly used motion capture systems and their composition characteristics,advantages and disadvantagesThen it finished collecting,classifying and reorganizing the application research achievements using motion capture techniqueAfter that,it presented a systematic review of the achievements from many aspects,including the intangible cultural heritage protection,digital simulation training and teaching,movie and game animation,body posture research,ergonomics research,etcFinally,it proposed that the technique should be applied to some new projects based on analyzing and summing up the achievements

Overview of location privacy protection technology

Abstract: As the proliferation of mobile devices such as smartphones and tablets,location-based services(LBS) are becoming increasingly popular.LBS,albeit useful and convenient,posed a serious threat to users' privacy as they are enticed to reveal their locations to LBS providers via their queries for.How to protect users' privacy against potentially compromised LBS providers is of vital importance to the well being of the LBS ecosystem.Some solutions have been proposed to deal with location privacy problems in location-based services,however,there is still a distance to be widely used.In order to resolve the contradictions which have been not solved,and deeply study with the related topics,this paper overviewed the related literatures from four aspects: the not k-anonymous location privacy technologies,the k-anonymous location privacy technologies,the k-anonymous location privacy in the P2P environment technologies,and the anonymous location privacy for continuously querying technologies.It introduced the related algorithms.Finally it concluded the problem remaining and development orientation of the location privacy protection technologies.

Analysis of user influence in microblog based on individual attribute features

Abstract: To improve the accuracy of key-person recognition in microblog topics,this paper proposed a new method of user influence analysis named PBF which based on individual attribute features.The method firstly used information diffusion characters to measure user influence,and then made regression analysis with individual attribute features to find out the ones effecting user influence most,with which predicted user influence.The experimental results indicate that the recognition efficiency of PBF method is obviously superior to RNF,improving the accuracy of key-person recognition effectively.

Survey on privacy preservation in IoT

Abstract: By establishing the architecture of IoT,this paper analyzed the privacy threats faced by sensor layer and process layer of the architecture in detail.It summarized the current privacy-preserving methods associated with IoT systemically,especially for methods of anonymization,encryption and routing protocols.At the end of the paper,it pointed out future research directions on privacy-preserving for IoT.

Gesture extraction and recognition research based on Kinect depth data

Abstract: Aiming at the problem that gesture recognition technology based on vision required a lot on environment and background,this paper presented a gesture extraction and recognition scheme based on depth data.It utilized Microsoft Kinect to capture gesture depth map,converted the depth map to 3D point cloud,and then used depth map filter to obtain gesture data.After direction adjustment for the gesture,calculated and imported the interval distribution feature of gesture depth information to support vector machine for training,thus implementing gesture recognition for number gesture 1 to 5.Experimental results show that the average recognition rate of this scheme is 95%,and the scheme makes high precision with simple device and good robustness.

Study of logistics network optimization model considering carbon emissions

Abstract: Considering carbon emissions,this paper proposed a problem of logistics network optimizationIt established the distribution center selection and demand matching model considering carbon emissions and the distribution routing optimization model on minimum carbon emissions respectively in two stages,and used the software Lingo 90 to solve the modelsThrough using the model of the second stage in conjunction with the model of the first stage,some decision problems about the distribution center location,it solved the demand matching between different logistics nodes as well as the vehicle routing optimizationFinally,an actual example verified the proposed models,and compared the optimization results when considering and not considering carbon emissionsThe results show the proposed method of minimizing comprehensive costs of entire logistics process considering carbon emissions is more practical value than the conventional method of not considering carbon emissions

Task scheduling algorithm in cloud storage system using PSO with limited solution domain

Abstract: This paper studied the task scheduling of the cloud storage system.Firstly,it analyzed the differences between the cloud storage and the cloud computing,and pointed out the existing PSO based task scheduling algorithm of cloud computing could not ensure the solution was meaningful to cloud storage,namely some solve may ask nodes offer data that they didn't have.In order to address these problems,the existing PSO based scheduling algorithm was improved by limiting the initialization solution and the solution search space in the exist solution space which could ensure solves were meaningful.The simulation results show that the algorithm can save 77% iteration times and also save about 4 times execution time and offer meaningful solutions for cloud storage system.

Wavelet entropy used in feature analysis of high speed train bogie fault signal

Abstract: Performance monitoring and fault diagnosis for the critical component of bogie is very important. Simulation data of high speed train bogie fault signal was selected in the data experiment. Based on multiresolution analysis,wavelet entropies were extracted to reflect the complexity level of the vibration signal on scales. In the high dimension composed by several wavelet entropy features,the dates from four fault patterns were classified and recognition rate is above 90% when the speed over 200 km / h. The wavelet entropy feature is effective for fault signal analysis of high speed train bogie.

Improved fully homomorphic encryption over integers

Abstract: The efficiency of the present fully homomorphic encryption scheme is extra low and far from practical application.How to improve the efficiency and security of fully homomorphic encryption become the focus and pitfall of academic research.In order to improve the efficiency,this paper put forward an improved scheme of the base of Dijk's scheme using Gentry's fully homomorhpic technology.The improved scheme could encrypt 2 bit plaintext each time and reduce the public key size to(λ7),accordingly the improved scheme was more efficient than Dijk's scheme and had smaller public key size.The security of the proposed scheme was based on both the approximate GCD problem and the sparse-subset sum problem.

Threat assessment based on entropy weight grey incidence and D-S theory of evidence

Abstract: By combining entropy weight grey incidence with D-S theory of evidence,this paper proposed a method of threat assessment to deal with the uncertain information of aerial threat target.It employed the theory of entropy to acquire the weights of different indices.Meanwhile,it determined uncertain degrees corresponding to different indices through the methodology of grey incidence.Subsequently,it obtained the Mass functions of different targets in different indices.It carried out the fusion of different Mass functions on the basis of D-S theory of evidence,sorted the targets according to the belief function value.A numerical example demonstrates that the aforementioned method is reasonable and effective.

Overview on label propagation algorithm and applications

Abstract: This article introduced the theoretical study of label propagation algorithm,analysed its characteristics and summarized its applications in multimedia information processing,retrieval,annotation,classification and community discovery,etc.Finally,this paper proposed the future prospects and the trends of developments of the LPA algorithm.

New algorithm for similarity calculation of Chinese character glyph

Abstract: In order to effectively identify similar Chinese characters by using the character glyph description technique,this paper proposed a method for similarity calculation of Chinese character glyph based on triple recursive representation.Firstly,it represented Chinese characters as triple: Chinese characters structure,the first part of Chinese characters and Chinese characters tail parts.It described Chinese characters as the prefix expression with the character components for the operation object,character structure for operator.Secondly,through establishing the recursive model of similarity calculation of Chinese character glyph,decomposed the calculation process into atomic components similarity comparison,and effectively reduced the computational complexity.Finally,it used to calculate the best similar Chinese characters.Experimental results show that the proposed method has a high coincidence with human perception.The algorithm is feasible and effective.

Prediction models of electronic products design defects based on rough set Bayesian network theory

Abstract: This paper presented an integrated method based on Bayesian network theory and rough set theory to analyze and predict the electronic products design defects.During the process that establishes the relationship among influence factors of electronic products design defects,it is usually difficult to establish directly the Bayesian network prediction model.This paper established the network topology and nodes conditional probability tables by using rough set theory.Then obtained the electronic products design defects prediction model with Bayesian network parameter estimation.At the end of this paper,the Bayesian network model of electronic products design defects,which is proved to be efficiently,is established as well as the learning and reasoning method.

Image encryption scheme based on hyper-chaotic sequence

Abstract: In order to improve the security of secret image transmission,this paper put forward a kind of image encryption scheme based on hyper-chaotic sequences.Firstly,it added plaintext pixel information to the transformation of each element of hyper-chaotic sequences which resulted in both key sensitivity and plaintext sensitivity.Then,it shuffled and encrypted the plaintext pixels according to the separation scrambling key sequences and the gray replacement key sequences pixel by pixel.The simulation results show that the gray distribution of encrypted image is balanced with its cipher-text sensitive to plain-text and key,and it is secure with strong resistibility to various attacks,such as statistical,exhaustive and differential attacks,verifying the feasibility of this scheme.

Survey of image segmentation methods based on Markov random fields

Abstract: This paper presented and discussed all the typical approaches based on MRF.First it provided a general framework of MRF in image segmentation and new progress for MRF modeling in recent years,and then emphatically presented the major image segmentation algorithms based on MRF,including graph cut algorithm,normalized cut algorithm,belief propagation algorithm,etc.Finally,it pointed out the future work for these segmentation algorithms.

Moving object detection based on consecutive blocks frame difference and background subtraction

Abstract: This paper proposed a novel approach for moving detection,which employed blocks frame difference and background subtraction.It used block processing method to establish the initial background model and it divided the sequence ima-ge of video into several blocks which was detected by self-adaptive threshold inter-frame difference,and separeted roughly the motion region.Then it carried on fine-grained segmentation through double-threshold background subtracting for the motion region,and used the adaptive background updating to overcome the light changing and background interference.Experimental results indicate that the method has the characteristics of fast operation and great robustness,and it can detect the moving object effectively.

Research on energy saving algorithm of datacenter in cloud computing system

Abstract: This paper briefly described the definition and characteristics of cloud computing,and focused the high energy consumption problem of cloud datacenter.Based on the classification of the energy saving algorithm,it researched the three kind of energy saving algorithms in a big way,including the energy saving algorithm based-DVFS,energy saving algorithm based-virtualization and the energy saving algorithm based-turn off/on of hosts,and made a comparative analysis of the advantages,disadvantages and applicable scene of these algorithms.Finally,it sumed up the further research problems for energy consumption management in cloud computing datacenter.