Abstract:Deep learning has become a popular research direction recently and has been applied to many fields. But the performance issue caused by the complicated learning model and massive training data has becomes an obstacle of deep learning evolution. As the development of processor technology, the core number and performance of the processor have increased rapidly. However, the acceleration research is restricted mainly due to the low parallelism and high memory consumption of the training algorithm. The paper introduces the background of deep learning and its training algorithm, then summarizes current deep learning acceleration works. Further, it analyzes classic deep models training algorithm to explain the causes of deep learning performance problem. Based on the analysis, it lists the challenges about deep learning evolution and makes proposal to address them.

Abstract:Crowd analysis becomes the most active-oriented research and trendy topic in computer vision nowadays. Within the crowd, there exist many behavior anomalies or abnormalities. There are many ways of detecting these abnormalities such as crowd density estimation, crowd motion detection, crowd tracking and crowd behavior recognition. The abnormal behaviors of crowd are analyzed, and the three key features of abnormal crowd are summarized. The feature extraction, anomaly identification technology, anomalies classification and databases of the crowd are respectively summarized, and the current problems, as well as the suggestions and comments about the future direction of research are then proposed. Article has a certain reference value for the research in related fields.

Abstract:The popularity of mobile system has attracted malware developers. To ensure the security of users' privacy and property, there has been numerous solutions targeting current security issues of mobile system. In this paper, it introduces the prevalent mobile operating systems and the mobile application security analysis methodologies. In addition, it makes analysis and discussion. At last, it proposes the prospects of possible directions of mobile application analysis approaches and endswith a conclusion.

Abstract:In the safety-critical systems, potential hazard may lead to catastrophic accidents, therefore it is greatly significant to analyze its influence. There are two causal logic relationships from risks to accidents that are deterministic and non-deterministic relationship. The former can be analyzed by event tree and fault tree. However, due to the latter's uncertainties and lack of data, Bayesian network could be applied as an appropriate tool. However, it is a difficult task to determine the conditional probability table. This paper proposes a new allocation method of conditional probability based on fuzzy logic. Combined with the risk model of influencing factors established, Bayesian network is used for quantitative analysis. Finally, the feasibility of proposed method is evaluated through case study.

Abstract:Mobile intelligent terminal platform combines telephone communication, social contact, online shopping and many other functions in one, the destruction of related services by malicious software could undermine the safety of personal privacy and property. Remote attestation, which is a core function of trusted computing, provides a scheme to allow mobile intelligent terminal to assure the secure running state to remote service provider. Traditional remote attestation model is mainly used in computer platform, which is unable to adapt to the environment of smart device where software updates frequently and works with many different service running together. To address on these issues, this paper presents a new behavior-based attestation model for smart device. Way from software developers to define behavior limitation, OS to enforce the limitation, service provider to custom strategy to attest smart device, successfully meets the requirement that corresponds service and should run in a secure environment. The implementation and evaluation of the model prototype shows that the model has strong security capabilities as well as efficiency.

Abstract:In recent years, with the rapid development of the Internet and the Internet of Things, the mass of data in many applications will appear, and a large part of the data is in the form of streaming data. Streaming data is characterized rapid, massive, disorderly, and also requires quick response. The research shows that the traditional relational database is not suitable for the application of the streaming data. Therefore, how to develop a new type of data query system is a hot topic in the current research. In this paper, it uses the advantages of several prevalent data management systems, analyzes the key problems of the streaming data query system and consideres the definition of data interface, data preprocessing, query language definition, query execution process, system monitoring, system interface and other issues to design and implement an available streaming data query system. Finally, it evaluates the system by the specific news streaming data. The results show that the streaming data query system has a good data query performance.

Abstract:As to the vast video information and its loose distribution, a vertical search engine is especially designed for video information through lucene and heritrix. With BKDRHash algorithm, a pointed video research is implemented. Web pages capture, content filtering and index set-up are integrated to greatly decrease the consumption the system, changes the traditional method to capture the whole page by the search engine, filter the useless information. It is creative in picking up page information and picking process of playing linking.

Abstract:Since the traditional web service identification methods are highly depended on the differences of server software implementation, unfavorable to expand to other terminal devices, a kind of proactive identification framework with double factors for cyberspace terminal device is proposed. The framework analyzed the network fingerprint characteristics of terminal device from perspectives of banner and web fingerprint. Firstly, banner information of web service in terminal device is mined from “Server” field and “WWW-Authenticate” field and HTML source code in HTTP packet header; secondly, web fingerprint of terminal device wasis extracted from image files and text files; meanwhile, it analyzes the effectiveness as well as advantages and disadvantages according to the specific cases. The results show that the framework solves the problems of extracting banner information simply and being limited by the differences of server software implementation in traditional identification methods. Finally, it summarizes research results and discusses future improvements of the framework.

Abstract:Power grid data has become an important asset of the enterprise, but power grid enterprises lacks effective technical means to solve the grid data assets life cycle monitoring and management. From the design principle of component, loose coupling and standardization, the power grid data assert management platform based on SOA is designed and realized. It designs a method of capturing the database log to realize real-time acquisition of data assets information change, and account management, management of the panoramic view, daily monitoring and management, traceability management, valuation management and process management data asset management the core function, and realizes the standardized management of grid data assets, and shows the status of data assets from the business, technology, management and other perspectives.

Abstract:In this paper, based on iOS system as the platform, the chronic disease tracking client is applied to the intelligent mobile terminal equipment that can download the scientific exercise prescriptions from the health cloud. These exercise prescriptions are generated by results of hospital physical tests and questionnaire survey data. And using a wearable device, the client can obtain user's physiological data via Bluetooth 4.0 protocols, and acquire user's motion data combined with apple mobile device built-in coprocessor, so as to achieve the purpose of chronic disease intervention scientifically and effectively.

Abstract:Mind map as an efficient visualization tools, its use of online and offline teamwork is also more and more, but the use of online collaboration tools is not much.The paper describes an implementation of a teamwork tool adopting B/S structure, whose significant function moduleis Mind Map. In this paper, the main functions, system structure, key technology and the related engineering method in the development process would be stated. It also talks about how to use TDD-Test-Driven Development and the IOC Vessel to optimize the favorable code organization structure and the ability of extension, maintenance and secondary development would be provided. Contrasting with the linear task management in the traditional collaboration system, the mind map is the visual and nonlinear tool aiming to stimulate thinking as well as brain bloom and flow and would have better interaction and lower threshold if we use the mind map to reveal the mutual relation of the teamwork tasks, and manage the project in the form of graphical interaction.

Abstract:With the development of robotic applications, service robots get into people's lives. But the robots themselves have limited computing capability, especially in processing the complex information received from different sensors. Thus it is not enough for a single robot to be competent to deal with the complex scenes. This paper proposes a cloud robot computing framework (CRCF), which combines the smart home facilities nearby and powerful hardware servers in the clouds. CRCF system platform is designed to take advantage of cloud's computing large data capability to extend the computing power and storage capacity of the robot. At the same time, CRCF system combines with third-party services and hardware to extend the sources of information and service functions of robots. Finally, through a remote control robot experiment, it verifies the CRCF system platform's capabilities and performance in combination with intelligent hardware and third-party cloud applications.

Abstract:The big data computing is one of the researches focus in the Cloud Computing nowadays. With the development of electric power information, the business complexity continues to increase and the amount of data is increasing quickly which makes the traditional way of computing be worse and worse. This paper provides an effective computing of big data base on the Objectification Parallel Computing (OPC) to solve the above challenges. The small data set is split from Big Data, is distributedly stored in memory of OPC Cluster. In the effective compute server, making use of the thought of divide and rule and tree merging, there are two stages. The first stage is local data calculate. The intermediate calculation result can be obtained. The second stage is multistage summarizing. The final result can be returned to user. The solution has been applied to the power production management system (PMS) of State Grid of China. The results show that solution is efficiently reliable and meets users' requirement.

Abstract:With the development of intelligent terminals and mobile internet, smart phones become more and more popular. Consequently, applications on them also become more and more diversified. Developers are facing a big problem of how to develop intelligent terminals' applications more quickly and efficiently. This essay analyzes and studies currently existing developments of cross-platform' hybrid application, based on which a hybrid applications' runtime on Android Platform is designed and implemented. Taking advantage of this runtime, developers can develop Android applications only via HTML, CSS, JavaScript. It brings much convinces to developers. This runtime environment is compatible with hybrid applications' runtime on desktop system. Applications developed via this runtime can be operated on desktop system after different screen adaptation.

Abstract:With the emergence of high-performance smart phones, mobile Internet is becoming mainstream, a large number of applications transfers from PC to mobile terminal, Convenient office also brings a variety of security risks and management issues. This paper studies the enterprise mobile application sandbox Technology, it realizes enterprise mobile application and data and personal applications and data isolation, in order to protect the security of enterprise applications and data. It has been proved through practice, based on the sandbox technology to develop products to meet the requirements of enterprise data protection, and it is convenient to use.

Abstract:In order to improve the efficiency of quality data acquisition in production field and to meet the needs of mobile data collection, this paper proposes a field quality data acquisition system based on android platform. The system uses tablet or smart-phone as client to communicate with server via WIFI or 3G/4G to realize mobile data collection and uses the user-interface configuration files to realize the customized quality data acquisition interface which makes it have good flexibility. Using the rule configuration files to realize the business logical control makes the client have a certain amount of intelligence.

Abstract:In this paper, a new kind treadmill system based on Kinect is developed, and running action recognition algorithms are presented. This system is under the development of Unity game engine, 3DSMax, Maya, Photoshop modeling and graphic design software, it obtains various skeletal data of human running action by using Kinect- Microsoft somatosensory equipment and personal computer, skeletal binding based recognition algorithms of running, waving, jumping and squat were given under analysis of movement feature. Interactive running fitness entertainment sports can be implemented in this system through the recognition of human actions. The experimental data shows the effectiveness of action recognition algorithms based on skeletal binding, the treadmill system has the advantages with smaller volume hardware equipment, better software operating method by using human body action instead of mouse and keyboard, and also combines function with the running fitness and entertainment, replaces the traditional treadmill effect in some parts it has the very good somatosensory interaction effect and practical value, the research results can be used for more human-computer interaction application areas.

Abstract:The comparisons and synchronizations of remote atomic clocks' data are vitally significant for accuracy and stability of atomic time scale. This paper introduces GPS common-view method for calculating the time difference between remote atomic clocks. The design and implementation of remote atomic clocks' data exchange system by file transfer protocol is detailedly discussed.

Abstract:High energy particle accelerator is necessary in the research of modern high energy physics, software of beam dynamic simulations is of great importance. This paper introduces the development status of a 3D heterogeneous parallel beam dynamic simulation solver, NEWBEAM-MIC. The purpose of this research is to speed up the beam dynamic simulation software and help to realize design and optimization of the accelerator by latest heterogeneous supercomputers. The motion of the particles through DTL and SOLENOID devices are simulated by this software. This software is called NEWBEAM-MIC, which is developed on the foundation of previous CPU version, NEWBEAM-CPU. It assigns simulation work of particle tracking on MIC, thus accelerates the procedure by taking advantage of MIC's multiple threads. Benchmarks verify that NEWBEAM-MIC can simulate 109 particles through DTL and SOLENOID devices on TIANHE- 2 using 100 CPUs and 100 MICs. It can achieve 100 times acceleration than the CPU version software, NEWBEAM-CPU. Considering using multi-threads on MIC, for the same number of particles, using 100 CPUs and 100 MICs with MIC's maximum threads (224) opening, NEWBEAM-MIC can achieve 10000 times acceleration than CPU version with only one thread. It proves that, the software developed in this paper can accelerate CPU version quite well by making use of the great power of the MIC supercomputer.

Abstract:For DBSCAN algorithm can't cluster some variable density data sets effectively, VDBSCAN solved this question by a k-dist figure to automatically obtain the neighborhood radiuses of various density levels of data objects. k-VDBSCAN algorithm obtains the parameter k automatically to reduce the parameter k ‘sinfluence in the final clustering results. Based on the data set itself, using the characteristics of the distance between the data objects, on the basic of the k-VDBSCAN algorithm, an clustering algorithm based on improved parameter k obtained automatically is proposed. Theoretical analysis and experimental results show that the improved algorithm can effectively automatically obtain the value of the parameter k and the clustering results and time efficiency has improved significantly.

Abstract:Repair in the distribution network is a very important part of the power system running, the lean of the distribution network emergency management can not only improve the quality of power supply service, but also reduce the economic losses of power companies. In this paper, a new fault forecasting method of repair in the distribution network is proposed. Firstly, based on historical data, the temperature, wind, the fault of the previous day, the maximum and minimum loads, etc are regarded as dependent variables. and feature mapping and preprocessing are performed on the variables. Then, the Random Forest algorithm is applied to establish the fault prediction model of the repair of distribution network, and to forecast the future failure rate in different regions, different power grids and non-grid faults and different voltage dimensions. The experimental results on the real power data show that the proposed method has better prediction efficiency and accuracy.

Abstract:To promote the standardization of hospital informatization construction work, strengthen the construction of hospital information management, supervision, guidance and assessment, this paper constructs a hospital information system evaluation system model with a broad representativeness, which makes the process of hospital information can be sustained and healthy developed. It calculates every single index weight by using the Analytic Hierarchy Process method, so that it can get the total score of the system performance testing and functional testing. With the score it can make the comprehensive evaluation of hospital information system. It scientifically evaluates the information system of two Affiliated Hospital of Xinjiang Medical University by the model, and gives the suggestions to improve the work.

Abstract:According to the characteristics of salt and pepper noise and the defects of the standard median filter algorithm, an improved median filtering algorithm based on support vector machine is proposed in this paper. At first, the algorithm filters the noise image by using median filtering and then defuzzification operations are performed on the filtered image. For those pixels in the defuzzified image whose gray scale values are the maximum or minimum value, support vector machine (SVM) classification is used to decide whether they are noise points. Finally support vector machine (SVM) regression is used to recover original signals from the noise points. Experiments and simulation analyses show that our algorithm can effectively remove salt and pepper noise, and has higher peak signal to noise ratio.

Abstract:Efficient and scalable metadata management system is the key to improve the overall performance of distributed storage systems. Traditional metadata allocation strategies result in load imbalance of metadata, and in the case of multi-process resource preemption, there exists low efficiency in response to user requests, limitation on the number of files stored and etc. The system cannot meet the demands of high-concurrency, low-latency. To solve these problems, a distributed metadata management system based on consistent hash and directory tree, is presented in this paper. It introduces a load balance algorithm for metadata migration which has low cost and fine granularity. Experiments show that this system can achieve the load balance of metadata, obtain low respond latency and effective distribution of metadata. and improve the scalability and availability of distributed systems.

Abstract:For the issues that the existing solutions for palmprint identification can't be very good to extract multi-resolution characteristic, this paper proposed a palmprint recognition scheme based on dual-tree complex wavelet transform (DT-CWT) and Levenberg-Marquardt (LM) neural network. Firstly, it convertsthe color image into gray image. Then, the region of interest (ROI) is extracted from the palm image, and constructed a histogram. Then, this scheme uses DT-CWT for 6 layers of wavelet decomposition and obtains the characteristic coefficients, and calculatesthe maximum value, average value and median value of the characteristic coefficients respectively. Finally, it uses LM neural network to make recognition and classification of palmprint. Experimental results on CASIA database show that the recognition rate of the proposed scheme has high recognition rate and lower recognition time.

Abstract:A fuzzy comprehensive evaluation model based on AHP is proposed to enhance the evaluation work, improve the level of company's management and reduce unnecessary duplication works. In this paper we firstly design a general internal evaluation index system according to the evaluation objects, as well as standards and processes. Then, it uses AHP to determine the weight of each index item, creates a dynamic evaluation model with fuzzy comprehensive evaluation method. Finally it develops an intelligent comprehensive evaluation system, as an effective complement to other business systems in power enterprise that provides a full range of automatic support to the appraisal management for the internal departments.

Abstract:Matrix Multiplication is a core computation in many applications, it has been observed that only a very small fraction of the matrix elements gradually change in these applications, and the full recomputation costs a lot, thus, incremental processing is a promising approach to handle this problem. This paper proposes a design of incremental matrix multiplication based on MapReduce with a fast method to capture the changes of the matrix, by using the abstract informationof matrix elements to calculate the changes of the matrix effectively and transforms this problem to series of equivalent join problems, which realizes a efficient matrix multiplication in an incremental way. Experimental results show significant performance improvements of our method compared to full recomputation when the change ratio is small.

Abstract:JADE algorithm is an improved algorithm of basic differential evolution algorithm (DE) with better convergence speed and optimization performance, whose self-adaptive parameter adjustment mechanism improves its global optimization ability. In this paper, we use self-adaptive differential evolution algorithm (JADE) for clustering and propose a new automatic clustering algorithm based on JADE，named as AC-JADE. Firstly, it takes double crossover strategy for clustering. Specifying to the encoding mode of DE used for clustering, it adds a new crossover strategy after the conventional two point crossover operation. This new crossover strategy acts directly on two clustering centers derived from parent vector and trial vector separately. Secondly, it makes improvements on the drawbacks that the selected clustering centers may deviate from the data set or they are too close results from the randomness of mode for choosing clustering center. Sifting clustering centers before choosing some of them for clustering results has a better effect. The experimental results carried on 4 UCI datasets verifies effectiveness of the proposed algorithm.

Abstract:The advent of the era of the Internet causes the “excessive” of knowledge dissemination, knowledge fragment and disorganization makes users who have requirements of a systematic study difficult to know how to start. How to realize personalized knowledge recommendation for users is one of the key problems to be solved in the intelligent library system. This paper uses the context preference extraction technology to obtain the users' interest. And it introduces weighted factor based on the intelligent library system as an example in the user-based collaborative filtering recommendation algorithm and the collaborative filtering recommendation algorithm, it better solves “cold start” in the recommendation to new users as, implements the service of personalized recommendation results.

Abstract:To solve the problem of graph data distributed processing, graph partitioning is more and more important. This paper studies the method of partitioning a large-scale dynamic graph. According to the characteristics of graph structure and dynamic graph, a dynamic graph partitioning algorithm based on neighborhood is proposed and implemented. The algorithm consists of two stages, static partition and dynamic adjustment. A static graph partitioning combines the cut-edges and the existing partition tools to get an optimized method. The dynamic adjustment focuses on the transferring of vertexes to its neighborhood. It calculates the value of vertexes transferring load and optimize the operations. Experiments on a series of common graph datasets show significant effect on the indexes of balance and cut-edges, which means the partition is reasonable and balanced.

Abstract:There has been high risk on internet financial P2P lending platform. In order to assist investors to obtain a better loan proceeds, this paper comprehensively takes the risk of bad debts, failure of bidding, interest rate and investor risk preferences and other factors into consideration, and proposes the investment decision-making algorithm IDEA (investment Decision Analysis): building the investor-loan network, taking advantage of investment behavior of loan information in the network to measure the risk of bad debts and make use of the graph based semi-supervised learning method to metric the risk of failed bidding to provide the basis for investment decisions. Experiment results on real datasets show that, compared with existing algorithms, our algorithm can not only get a better return on investment, but also assist investors with different risk appetite to make investment decisions.

Abstract:In order to enhance the immersive realism of motion blur in virtual reality system, an improved post-processing algorithm of motion blur based on Graphics Processing Unit (GPU) is proposed. Depth value of full resolution texture maps in before and after frames of scene rendering are stored in the depth buffer in this algorithm. It uses its depth information and mapping matrix of the current frame to extract the world coordinate of the object, and gets the velocity vector. Then along speed vector it samples Gaussian filters. Fuzzy factor is introduced and fused weighted linear sample values to achieve the most realistic motion blur effect. The results show that: the proposed algorithm withgood real-time and realism effect will not only not affect the fluency of rendering large virtual reality system, but also solve over ghosting or streaking.

Abstract:Label Propagation Algorithm (LPA) is a fast and efficient community detection algorithm and this algorithm does not need to know the prior information such as the number of communities. However, this algorithm has a large number of randomness, which leads to unstable results. In order to improve the stability of label propagation algorithm, we propose an improved label propagation algorithm (LPAMP). The algorithm is divided into two phases. In the first phase, vertices are clustered roughly by optimizing the modularity greedily; in the second phase, the labels propagate through the network based on the result of the first phase. Experimental results show that the proposed algorithm not only reduces the randomness of the label propagation algorithm, but also improves the stability and increases the accuracy.

Abstract:In regard to the energy efficiency in the Wireless Sensor Network (WSN), a kind of multi-type sensor scheduling algorithm on cluster structure is proposed in this paper. Both the health status of the nodes and the information quality of the nodes' transmission are considered in the algorithm. According to different health status of each node, the proposed algorithm adaptively selects task node. Reasonable sensors scheduling promotes the balance between energy consumption and task allocation, and ensures that the network runs smoothly. The simulated experiment result shows the good effectiveness of the proposed algorithm.

Abstract:The development of “English Chinese - Thai Translation Audible Corpus”, in the field of Thai public opinion analysis, it solves the problem of the lack of training corpus in dictionary segmentation algorithm. In this paper, the computerized information processing technology is used to organize, standardize, process and store large amounts of collected Thai corpus, and then more than 8000 of the Thai vocabularies are finished. And then it uses the dictionary translation and manual calibration to align corpus. Finally, Thai corpus are classified and marked normally, combined with syntax characteristics and syntactic semantic features of Thai language, and then the database of English Chinese - Thai corpus is constructed, which contains about 50 thousand of the query terms.

Abstract:The Simulation drilling system provides a good service for the enterprises and the emergency departments at present. However, it still has quite a few drawbacks. Such as the lack of modern graphic pipeline support, the excessive computation of scene rendering on the condition of multiple lights, the poor rendering effects caused by that it cannot make full use of hardware etc. In order to solve those problems, this paper is designed and implemented with a deferred shading pipeline for simulation drilling system. This pipeline is based on the Blinn-Phong lighting model, which builds a compressed G-buffer layout and uses Light volume to model the light source. The final experiment indicates that under the multiple lights circumstances, the deferred shading expressed in this paper will be more efficient than the conventional forward shading.

Abstract:Farmhouse is the emerging form of tourism and leisure, as modern farmers provide a leisure entertainment of returning to nature for relax and cheerful spirit to the cities Shangluo city life rhythm is slow, the surrounding environment presents the pastoral scenery, air pure and fresh nature, which can relieve the stress of modern, and is very suitable for farmhouse. This article discusses the fact from the rural environment, health, services, features four factors, use the analytic hierarchy process (AHP) and adopts the combination of qualitative and quantitative way to study the present situation and the influence of shangluo farmhouse, this paper, proposes some approaches and strategies to help create good, brand of farmhouse.

Abstract:The real-time parameter is a key point in PROFIBUS field bus protocol's performance evaluation, so how to analyze and count its communication parameter is of great importance. According to the problem, based on the detailed analysis of PROFIBUS field bus access protocol, a type of testing platform for PROFIBUS-DP communication performance is designed. Through this testing platform, these real-time performance parameters are analyzed including bag information rate, transmission efficiency, network average utilization, network throughput, transmission delay, token rotation time etc., and the relationship between bus cycle time and master station or packets number are also given out. So, the real- time performance of PROFIBUS-DP is quantitativly analyzed.

Abstract:With the basic function of CAD software system it's convenient to realize information analysis and pick up of built-in geometrical elements, but it's difficult to non built-in geometrical elements, it is intended to discuss how to build software system under HOOPS/ACIS platform, and how to deal with the problem of CAD's file model process, explainhow to decide the module processing system. And in details, achieve the function of main module, and the way of technical implementation.

Abstract:If you want to keep and get more mobile applications user resources in Mobile Internet Era that is the era of rapid iteration, you need to provide the best user experience. Although the mainstream mobile application is Native App, but Web App broughs new experiences from the development of low cost, fast iterative cycle, cross platform with the perfection of HTML5. Web App will replace Native App? How will the future of mobile application development? That is focus of mobile application. In this paper, we study the development trends of Native App and Web App in mobile application fields in their characteristics, operating efficiency, iteration period, development pattern, user experience and so on. Moreover, the paper also analyzes the overall development trend of mobile application.

Abstract:To offset the shortage of knowledge retrieval methods for product design process, while to reduce the blindness of knowledge push due to uncertain demand, a method according to users' interest feature to achieve knowledge push is proposed. Firstly the theory of design knowledge push based on users' interest feature is expounded, and the mothod of knowledge organization and management is given. Then the method for establishing user interest feature set is analyzed, and the match between knowledge feature and interest feature is realized by utilizing similarity matching algorithm. Finally case is given to illustrate feasibility and reliability of knowledge push method.

Abstract:In view of the current way of artificial erase the blackboard, this paper puts forward a new way of intelligent erasure. The system takes blackboard as the control object, the interactive interface as the terminal, and uses the embedded technology and stepper motor control technology to realize the collection of blackboard images, extraction of the blackboard coordinates, transmission of serial port data and control of stepper motor. The experimental results show that this system can clean the blackboard automatically, and has the advantages of flexibility, high integration, and convenience etc.

Abstract:Componentization can make the structure of the application system more clearly, simplify the development, deployment and upgrades of large-scale systems. However, the traditional methods of componentization have some drawbacks like high cost, poor system scalability,thewaste of resources and so on. In order to solve the above problems, an application framework based on microservice is realized to functionally decompose the application into a set of collaborating services. Using this framework, developers just need to focus on the business function development, and service registration, service discovery, access and monitoring is done by the framework. This framework can simplify the development; reduce costs and risks of code changes, testing, packaging and deployment; enhance capacity of dynamic need expansion; improve fault inside isolation.，enhance the fault tolerance of the system.

Abstract:Based on the traditional method of binaryzation, this paper divides the signal sequenceinto into several areas anddomulti-scale binaryzation process on the sequence. Under the condition of without increasing the number of symbols, it can get more meticulous and more accuracy complexity. After the experiment, according to our analysis, with the decrease of the calculating scale, the amplitude of the text's complexity increasing is much smaller than the image's. Acording the conclusion, the difference between the Lempel Ziv complexitys of different scales can be regarded as the feature of image and text. Therefore, it can apply the improved LZC algorithm to distinguish the image and text.

Abstract:Virtualization is the key technology of cloud computing. Hypervisor provides an abstraction layer between the virtual machine and the host hardware, which allows the user to allocate memory for running the virtual machines over the available memory of a host. This technique is called memory overcommitment. In order to reduce the impact of this technology on the performance of the virtual machines, hypervisor must provide an efficient memory reclamation mechanism. In this paper, the author proposes a solution: using the nonvolatile memory as a cache device for hypervisor to exchange page file data. The author divides the space from the system memory to simulate the nonvolatile memory devices, and modifies the algorithm in the KVM module, and sets five kinds of testing environments. The experiments show that compared with the existing Ballooning technology and Hypervisor swapping technology, using the nonvolatile memory and with low priority queue algorithm, the virtual machine performance can be improved by 30% and 50%.

Abstract:For watching videos from different equipment manufacturers on the unified interface, a web video surveillance plugin is designed. According to the network configuration, the plugis connected with the video device directly. Based on integration of the APIs which are provided by different video equipment manufacturers, the plugin is compatible with video devices from different manufacturers. Based on the mechanism of ActiveX and NPAPI, the plugin interface is developed. Through the plugin in different browsers, we can watch the real-time video and the local video, download and play the remote historical video. It provides the same remote video surveillance experience for users. The plugin has been successfully applied in the wisdom of park construction project of Shandong Runxing Chemical Technology Co., Ltd.

Abstract:This article is in order to solve the existing problems in the central air-conditioning system, such as energy efficiency information scattered and hard to be shared, designs a central air-conditioning energy efficiency monitoring system. According to the existing energy efficiency monitoring means and efficiency calculation methods, builds an energy efficiency evaluation system which is combined with the central air-conditioning diagnosis ontology library, to display energy efficiency monitoring results and diagnostic decisions for the central air-conditioning system devices and subsystems. This platform can provide convenience for energy efficiency monitoring of the central air-conditioning, it can also provide conditions for further maintenance for energy saving.