Traditional classifier system is considered as a multi-population GAs architecture which represents individual learning. In real world, people learn not only from individual experience but also from others. By introducing a concept called ‘social learning bonus’ an extended classifier system which mixed individual learning and social learning is implemented in an artificial stock market. The results suggested social learning leads to different market statistic property around a threshold, and a state like HREE may be realized endogenously.
This paper proposes an integrated approach to automatic information extraction for Forums, Blogs and News web sites using wrapper. This paper presents a tree alignment and transfer learning method to generate the wrapper. The tree alignment algorithm is adopted to find the best matching structure of the input web pages. A kind of linear regression method is employed to get the weight of different tag-matching. For wrapper maintenance, this paper presents a method using a log likelihood ratio test for detecting the change points on the similarity series which gotten from the wrapper and input web pages. Experimental results show that the method achieves high accuracy and has steady performance.
Mobile Agent technology has broad application prospects, mobile strategy of which will directly affect the performance of Mobile Agent and even completion of its Task. Therefore, the research on self-migration calculation of Mobile Agent has theoretical significance and application value. Sequence will be regarded as a part of selective task, so this paper only discusses migration issues of Mobile Agent under the case of sequential, selective and parallel. This paper proposes a new definition of the extended itinerary graph, and on this basis gives the mathematical calculation model of the fitness function, population selection operator, mutation probability of genetic algorithm. By improving the genetic algorithm, Migration Algorithm of Mobile Agent is designed.
A widely existing problem in contemporary web application servers is the delay of response time. A long response time will cause the loss of clients and the drop of commercial interest for web site operators. The web Quality of Service (QoS) control should be deployed in web application servers in order to meet different requirements from clients and applications. In this paper, we use classical feedback control theory to contribute a QoS guarantee, i.e. the absolute delay guarantee (ADG). ADG is designed to ensure that the average queuing delay for the requests with high priority is no more than the threshold configured. The approximate linear time-invariant model of the database connection pool (DBCP) in Tomcat web application server is implemented through system identification experimentally. ADG controller is designed to use Root Lacus method. According to the error between the measured queuing delay and the reference value, the controller is periodically invoked to calculate and adjust the probability for different classes of requests to use a limited number of database connections. All components of the closed-loops of QoS guarantee is implemented for HTTP dynamic requests in the database connection pool. In order to evaluate the performance of the closed-loop systems, we design some experiments and gain experiment results. Experiment results show that the controller we design are able to handle fluctuate workloads effectively. ADG can be achieved well in the DBCP in Tomcat Web Server even if the number of concurrent requests changes abruptly.
With the changes in computing environment, the research on mobile Agent has attracted wide range of applications; Migration strategy of mobile Agent becomes one of popular research topics. Based on extended itinerary graph, this paper presents encoding scheme and data structure of sequential, selective and parallel tasks, and the using of improved genetic algorithms to solve problems of migration of mobile Agent. The paper implements the migration stimulation toolkit of mobile Agent, in order to test the effectiveness of the proposed method, and can meet the requirement of migration strategy in accuracy and convergence speed.
This study is part of a larger project that ultimately aims to enhance effectiveness and relevance of academic libraries in today's ubiquitous and highly networked environments. The current study asserts that management of customer knowledge, enabled by appropriate customer knowledge taxonomy, will potentially lead to both enhancements in the current customer services as well as identification and design of new and innovative customer services in general, and in academic libraries in particular. As the first stage of the larger project, the present study provides a high-level methodology for development of a model warehouse as a precursor for attainment of the above higher goal using the theoretical perspective of knowledge management. Findings of the current study are expected to benefit knowledge-based organizations that may already have implemented some kind of customer relationship management system with an accumulation of associated historic customer data.
Redundant reader elimination is a critical approach to optimize the performance of large scale RFID networks. In this paper, an algorithm for redundant reader elimination (MRRE) is presented, which is based on EPC network architecture. MRRE algorithm eliminates redundant readers using the tag information in RFID middleware, with no need for “write-to-tag” operations. Simulation results reveal that, compared with the canonical LEO+RRE algorithm, MRRE can improve the detection rate of redundant readers from 6.27% to 20.80% and decrease the processed quantity of tags from 4.50% to 35.73%. Moreover, MRRE algorithm can advance the deployment rationality of RFID networks.
We tried to verify a low-cost eye tracking device(Model KSL-240) designed in our Lab, using a commercialized eye tracker device(Tobii 1750). Participants were asked either to move eyes or to move mouse cursors to the location of a red circle presented in the target display, and they were assigned to three experimental conditions(KSL-240/Tobii 1750/Mouse Tracking). As results, 1) accuracy of the Mouse Tracking condition was 100%, 2) accuracy of the KSL-240(92%) eye tracking condition showed better performance over the Tobii 1750(79%) condition, 3) mean response time of the eye movements in the KSL-240(598ms) condition was somewhat slower than that of the Tobii 1750(466ms) condition, and the Mouse Tracking (808ms) condition was the slowest. In conclusion, it was suggested that the model KSL-240 eye-tracker could be used as a device which might be able to substitute more expensive, commercialized devices.
It is expected that many heterogeneous wireless networking systems such as 3GPP Long-term Evolution (LTE) system, WiMAX/WiBro system and WLAN system will exist in the next generation wireless communication environment. Under such environment, various communication services will be introduced by integrating heterogeneous wireless networking technologies, especially at radio access networks. In order to provide such integrated communication services, integrated radio resource management and seamless vertical handover across such heterogeneous access networks should be supported. In this paper, to tackle these problems, we propose a Generic Link Layer (GLL) based Common Radio Resource Management (CRRM) architecture and three vertical handover decision algorithms. We propose a vertical handover decision (VHD) algorithm that is a combination of a policy and a multi-criteria decision making (MCDM) approaches designed for achieving effective and seamless handover between heterogeneous radio access networks, especially between the LTE and the WLAN systems. We also propose the Data Mining(C4.5) based VHO decision making algorithm not only to consider various VHO decision factors but also to decide the handover trigger point adaptively to the network traffic situation. In order to verify effectiveness of the proposed CRRM architecture and the three proposed VHO decision making algorithms, extensive simulation studies are performed under various traffic situations. Simulation study shows that the proposed two VHO algorithms outperform the conventional received signal strength (RSS) based VHO algorithm in terms of the data throughput, the handover success rate, and the system service cost.
Open environments, like the Internet and corporate intranets, enable a large number of interested enterprises to access, to process and to present information on an as-needed basis. These environments support modern applications, like virtual enterprise (VE), which is a network alliance of cooperating enterprises. VE, involves a number of heterogeneous resources, services and processes. This paper proposes a framework for virtual enterprise based on the multi-agent technology. The paper also discusses different types of the agent and their relationships in the model.
The way teams use virtual collaboration tools, such as wikis, email systems, social networks, or version control systems can provide indicators for the success or failure of projects. We previously created a platform that allows to collect and analyze these virtual collaboration activities during project runtime in a non-interfering manner. In this paper, we provide a formal definition of collaboration patterns to enable sharing of beneficial or detrimental collaboration behavior amongst scientist and practitioners. We further define a mapping from collaboration pattern descriptions to SPARQL queries that allows to automatically test other projects for occurrences of the described behavior. By that, we provide a research tool that is able to stimulate relevant and rigorous findings in empirical engineering research and lead to the creation of a shared repository of patterns that reflect best practices in virtual team collaboration.
This study proposes a method to estimate brain-surface deformation, which usually happens during cranium removal for brain surgery. The deformation estimation method utilizes a stereo-vision camera system to estimate brain surface structure and adopts thin-plate spline algorithm (TPS) to model the brain deformation. Before brain deformation, the full brain surface needs to be reconstructed by using pre-operative medical images or surface digitizing devices. Next, feature points on the surface are detected and reconstructed into 3-D shape from the images obtained by the stereo-vision cameras. While the deformation occurs, we estimate the corresponding points of the feature points in the current images. By associating the feature points before and after brain deformation, a globally smooth transformation can be estimated by TPS. The transformation is then applied to the full surface data, and the deformed brain surface is thus obtained. In the experiment, a porcine brain is used as the phantom to test the proposed method, and the result shows the average mean error is below 1.1 mm, which proves the superiority of this approach.
In rapidly changing software industry the organizations have to adjust their processes to meet the changes. Software process improvement (SPI) is a continuous effort in software organizations, which aims at improving the organization's processes and quality of products and services. However, an organization needs to either initiate SPI efforts or adjust them to organizational changes, such as growth, internationalization, mergers and acquisitions, and adoption of standards or process improvement frameworks. Often, the changes affect software processes and require making modifications to processes. Software industry is knowledge-intensive business and emphasizes human capital. Employees need to be motivated, aware, and trained to SPI in order to be capable and keen on improvement work. The research problem in this paper is how does training support software process improvement in organizational changes? We have analyzed process improvement and training activities in three different SPI and training related research projects within period of 1999–2011. The organizations' interests and reasoning towards SPI, improved processes, and trainings that support process improvement have been studied. Our findings show that training can be used to support SPI, to increase understanding of processes and process improvement, and to adjust organizations' SPI to changes.
Software development effort estimation is vital to project success. Both underestimates and overestimates of software effort are universal phenomenon in the software industry, which is critical for resource allocation and bidding. The Class Point (CP) system level object-oriented size measure was proposed as an adaptation of Function Point (FP) analysis for effort estimation purposes. However, the dataset for validation is limited to 40 university student's projects. This implementation is not only threatens the external validity of the conclusions but also the type of project has raised the issue on its weight allocation. In this paper, we proposed an alternative sizing approach for object-oriented effort estimation by integrating the FP's new calibrated weights into the CP's class complexity evaluation criteria. A preliminary correlation coefficient investigation on this integration using the six industrial verified object-oriented projects have shows that the proposed approach can be used with high confidence for effort estimation under object-oriented development paradigm.
To explore the intellectual structure of accounting disclosure research in the last decade, this study identified the most important publications and the most influential scholars as well as the correlations among these scholar's publications. In this study, bibliometric and social network analysis techniques are used to investigate the intellectual pillars of the accounting disclosure literature. By analyzing 74,742 citation of 2,015 articles published in SSCI journal in accounting disclosure area between 1989 and 2010, this study maps a knowledge network of accounting disclosure studies. The results of the mapping can help identify the research direction of accounting disclosure research and provide a valuable tool for researchers to access the literature in this area.
This paper reviews the cross-cutting concerns and determines why it is commonly observed. We then study the Aspect Oriented Programming (AOP). Two examples are discussed to show how to deal with the frame work w/o aspects. A general solution of handling framework with aspect is presented. We also discuss the cohesion and coupling and the reusability of the pattern implementation. We finally show and discuss what other researchers found in cross-concerns and aspect and potential AOP research areas.
Total quality management (TQM) is a philosophy and a system for continuously improving the services and products offered to customers through ongoing refinements in response to continuous feedback. With regards to higher education, quality education needs to be emphasized in ensuring progressive intellectual capacity building. Both quality assurance and management are critical in ensuring that higher learning institutions provide better services to their primary customers. The continuous improvement and growth focus of TQM would offer more excitement and challenges to students and lecturers as compared to a “good enough” traditional learning environment. In this paper, elements of an education system that focus on total quality assurance and management are discussed. The most important element is to ensure that each staff within the educational system has adequate awareness on quality education. A clear mission and vision must be in place and a proper learning system must be adopted. This paper emphasizes on the mastery learning approach, which is in line with TQM. The learning process follows the Plan, Do, Check, and Act steps or the PDCA cycle. It contains the following steps: Plan, teach (Do), Check (formative evaluation), revised teaching (Act), and Test (summative evaluation). The two main objectives of using TQM in education, which are to improve learning and to improve cost effectiveness, are emphasized. The paper concludes with a framework that can be used as guidelines to total quality assurance and management in higher education. It is hoped that this framework is useful in building intellectual capacity in higher education.
In recent years, more and more people use Internet and with the improvements of communication technology, information-sharing trends has surged over Internet. Bit-Torrent is a well-known P2P application used for distributing large amounts of data and exchange digital content with high efficiently and scalability. Unfortunately, many users do not get authorization but transmit files through Bit-Torrent illegally. The piracy problems become a very serious issue because the digital contents can be copied and redistributed easily through P2P network without license owners' authorized. For this reason, digital rights management (DRM) is considered as a general preventive measure to avoid copyright infringement. However, a conventional DRM system implementation on distributed P2P network would be quite difficult. Therefore, we propose an efficient DRM model on P2P network that to prevent violation of copyrights for digital content owners and to provide more attractive motivation for users.
This paper proposes a nevel approach supporting evaluation of software Safety Integrity Level (SIL) which is a relative target level of risk-reduction provided by a safety function using failure frequency. Software safety on embedded systems has become an important software engineering challenge, since the embedded system is closely used to human life and the software directly controls most operations of the embedded system. To address this challenge, several techniques have been proposed to analyze and evaluate the software safety. However, although these analysis and evaluation techniques have been proposed, still experts need to qualitatively evaluate the software SIL, even if the representative international standard for safety, IEC 61508, deals with software SIL evaluation, since existing standards enumerate only recommended software techniques for evaluating software SIL. Therefore, certification of software SIL highly depends on a third-party consulting company and this high dependency makes additional cost. Hence, our approach can be used without the dependency before formal certification of software SIL. Our approach identifies possible failures on embedded systems and calculates a probability of failure frequency using Markov process. Using the probability, we define quantitative measures to evaluate software SIL. We also conducted a case study using open data to evaluate our approach. The result of case study demonstrates that our approach can be a reasonable method to support quantitative evaluation of software SIL without expert's knowledge.
The increased need in promoting fitness activity and rapid growth of smart-phones has urged the development of mobile virtual fitness apps (MVFA). Yet, evaluation on these MVFA has been limited and, therefore, this study attempts to conduct preliminary evaluation on randomly sampled MVFA based on our proposed system workflow using theories of social support and persuasive technology, and the American College of Sports Medicine (ACSM) guidelines for exercise professionals. Results indicated the sampled MVFA mainly covered the stages 2 and 3 of our proposed ACSM-based training workflow. In terms of social support and persuasiveness of MVFA, the average scores of these two aspects were relatively low, thus resulting in a generally low average quality of coverage scores. Therefore, MVFA cannot replace human fitness trainers, but serve as assistant to trainers and trainees. Explanations and implications to trainers, trainees and MVFA developers are presented.
