294 Y. Sheng et al. results of FIRLowFilter component are sent to another Oscilloscope component to display. Users can compare the results of the Adder component with the result of the FIRLowFilter component in the two Oscilloscope components. (6) Execute step (5) iteratively until users push “stop” button. The experiment result is shown in Fig.11. The left frame is the signal of the Adder component, and the right frame is the signal processed by the FIRLowFilter compo- nent. In the experiment flow, we notice that the signals are changed according to the iteration of the step (5), which improves the effect of the experiment. Then user can get more knowledge about low-pass filter. 6 Conclusion In view of the question existing in the design and development of present VL, this paper proposed the design and realization methods of laboratory platform based on the integration of Java and Matlab. The combination of Beans components and Mat- lab in the VL-JM platform enhanced the developing efficiency of components. The VL-JM platform has the following characteristics: (1) Make the development of components efficient. (2) Be independent-platform. (3) Make it easy to construct one virtual laboratory. (4) Have friendly user interface. (5) Make the component reusable. References 1. JMatLink, http://www.held-mueller.de/JMatLink 2. Wang, J., Chen, S., Jia, W., Pei, H.: The Design and Implementation of Virtual Laboratory Platform in Internet. In: Proceedings of The First International Conference on Web-based Learning in China, August 17-19, 2002, pp. 169–177 (2002) 3. Cao, J., Chan, A., Cao, W., Yeung, C.: Virtual Programming Lab for Online Distance Learning. In: Fong, J., Cheung, C.T., Leong, H.V., Li, Q. (eds.) ICWL 2002. LNCS, vol. 2436, pp. 216–227. Springer, Heidelberg (2002) 4. Subramanian, R., Marsic, I.: ViBE: Virtual Biology Experiments. In: Proceeding of the Tenth International World Wide Web Conference (WWW10) (2001) 5. Wang, J., Lu, W., Jia, W.: A Web-Based Environment for Virtual Laboratory with CORBA Technology. International Journal of Computer Processing of Oriental Lan- guages 16(4), 261–274 (2003) 6. Muller, S., Waller, H.: Efficient Integration of Real-Time Hardware and Web Based Ser- vices Into MATLAB. In: ESS 1999 11th European Simulation Symposium and Exhibition, October 26-28, 1999. Erlangen-Nuremberg (1999) 7. Lobov, A., Lastra, J.L.M., Tuokko, R.: A Collaborative Framework for Learning Robot Mechanics: RIO-Robotics Illustrative Software. In: The 33rd ASEE/IEEE Frontiers in Education Conference, November 5-8, 2003, pp. 12–16 (2003) 8. Bai, Y.: Application Interface Programming Using Multiple Languages, March 21, 2003, pp. 266–287. Prentice Hall PTR, Englewood Cliffs (2003) A Virtual Laboratory Platform Based on Integration of Java and Matlab 295 9. Jianxin, W., Bei, P., Weijia, J.: Design and Implementation of Virtual Computer Network Lab Based on NS2 in the Internet. In: Liu, W., Shi, Y., Li, Q. (eds.) ICWL 2004. LNCS, vol. 3143, pp. 346–353. Springer, Heidelberg (2004) 10. Wang, J., Liu, L., Jia, W.: The Design and Implementation of Digital Signal Processing Virtual Lab Based on Components. In: Lau, R., Li, Q., Cheung, R., Liu, W. (eds.) ICWL 2005. LNCS, vol. 3583, pp. 291–301. Springer, Heidelberg (2005) 11. http://java.sun.com/products/javabeans/reference/index.html F. Li et al. (Eds.): ICWL 2008, LNCS 5145, pp. 296–303, 2008. © Springer-Verlag Berlin Heidelberg 2008 Multi-agent Framework Support for Adaptive e-Learning Wanjie Liang, Jianmin Zhao, and Xinzhong Zhu College of Mathematics, Physics and Information Engineering, Zhejiang Normal University, Jinhua, China wanjie.liang@zjnu.net, zjm@zjnu.cn, zxz@zjnu.cn Abstract. In the past years, agent technology is considered one of the most in- novative technologies for the development of software systems. Meanwhile, following the rapid development of Internet, particularly web page interaction technology, web-based learning has become increasingly popular. However, there not yet has a perfect framework in e-learning system’s software designing. This paper proposes a multi-agent framework to realize an adaptive e-learning system. Experimental results indicate that applying the proposed framework for personalized e-learning system is feasible and robust. Keywords: Multi-agent; Adaptive e-learning; Web-based tutoring. 1 Introduction In the past decades, the rapid growth of the internet has brought a great deal of changes in our educational environment. Internet and web rapid development has provided new mentality and method for e-learning. E-learning has a inherited merit like not limit by time, spatial and place in the traditional learning, learners may par- ticipate on-line study, on-line test, on-line discussion as well as on-line Q/A and so on. Otherwise it provides abundant, rich, colorful and an alternating interface between man and computer for study. It can stimulate learner’s study interest, thus the goal of acquisition knowledge, self-renewal even knowledge innovation is achieved. The merit of e-learning not only lies in it’s a very good content carrier could visit at any time, but also lies in it provides many exchanged channels allowed teachers and stu- dents to discussion. The emerging e-learning is reshaping the instructional community and provides tremendous cost savings for both instructors and learners[1,2,3]. But it lacks in interaction aspect, intelligence, personality, adaptability, simultane- ously and returning feedback information not in time, its easy to misleading in learner's study process. Otherwise, it is not realize leaner’s characteristics intelli- gently, it causes learner lose one's head when in front of the websites which filling lots of teaching information, and can not learning effectively. Therefore, how to en- hance the e-learning intellectualized degree is our urgent work. In recent years for achieved this goal, in the artificial intelligence domain, multi- agent technology provides a good opportunity. With the rapid development of AI(Artificial Intelligence),the agent technology is nearly mature. The agent has many Multi-agent Framework Support for Adaptive e-Learning 297 characteristics, such as autonomy, proclivity, reactivity, sociality, collaboration, intel- ligence, and so on. Thus, in the agent environment, educational application focus on information searching, information organization, scheduled events response, problem solving, knowledge mining and regular service of internet. Hence, using agent tech- nology, e-learning systems makes itself disadvantages up effectively[4]. In the real world, questions are extremely complex, individual agent function is extremely lim- ited, generally, it is very difficult to complete the assigned task, and then, needs to organize many agents to form multi-agent system through suitable system structure to undertake tasks together, the multi-agent system could make up the insufficiency of single agent and its function surpass single agent[13]. MAS(Multi-Agent System) technology has impressively emerged as a new para- digm for software development[5]. As autonomous software components, agents can interact through a standard protocol and collaborate with each other to achieve com- mon goals. MAS can help application designers to conceptualize solutions better: this paradigm may be more naturally suitable for certain types of applications; they can help improving code modularity and reusability; they can help hiding network, system and protocol heterogeneity. The features – autonomy, sociality, and communication possessed by agents make it easy to decompose a complex task into some simple ones and then assign them to individual agent that collaborate, negotiate and eventually achieve the common goal. Naturally, the agent-based software engineering paradigm is particularly suitable for developing various distributed systems because it could accelerate development with agent components and enhance modularity, speed, reli- ability, flexibility and reusability. At present, considerable research in agent technol- ogy applications for e-education has been conducted over the past few years. The main contribution of this paper is to propose a multi-agent framework to real- ize an adaptive e-learning system. In the e-learning context, the indispensable func- tions are the diagnosis(assessment),online helping, adaptive navigation and course- ware recommendation[6] and so on. In this framework, each function is undertaken by an intelligence agent. This paper is organized as follows: Section 2 describes the system architecture of the adaptive e-learning system and the functions of each agent. In section 3, we give a experiment and evaluation of the system. Finally, we draw our conclusion in Section 4. 2 System Architecture This section describes the system architecture, system components, and functions of each agent in the proposed personalized e-learning system. The system architecture of the adaptive e-learning system are outlined in Fig. 1. 2.1 System Architecture and Components Here, an adaptive e-learning system based on multi-agent technology, which includes eight intelligent agents, four databases and two repositories are presented. The eight intelligent agents are learner interface agent, teacher/expert interface agent, diagno- sis/assessment agent, adaptive navigation agent ,courseware recommendation agent, 298 W. Liang, J. Zhao, and X. Zhu auto-reply agent, courseware/testing items management agent and database manage- ment agent, respectively; the four databases include learner account database, learner profile database, testing items database and teacher/expert account database; The two repositories include the courseware repository and answer document repository. Learner nterface gent aims at providing a flexible learning interface for learners to interact with the personalized e-learning system. The eacher/ xpert nterface gent aims at providing a friendly management interface for teacher/expert to manage the courseware and testing items of the system . The function of atabase an- agement gent is to manage the four databases and the two repositories. All intelligent agents need interacting with the four database or the two reposito- ries must use the atabase anagement gent . The ourse- ware/ esting items anagement gent aims at managing the courseware and testing items of the adaptive e-learning. Teachers or experts can use the agent to create new testing items course units, upload testing items courseware to the testing items courseware database and delete or modify testing items courseware from the testing items courseware database. The uto- eply gent has two functions, one is to answer learner’s questions automatically, is to organize and manage the nswer ocument of the system. The iagnosis/ ssessment gent is the important agent in the system, which in charge of investigating the learner who first use the system, providing a final test while the learner finishes the whole learning process and storing these resulting information in the user profile da- tabase for personalized e-learning services. Moreover, the daptive avigation gent is responsible for guiding the learner’s learning process based on the learner’s study- ing situation and storing learning records into the profile database. The Fig. 1. The system architecture of the adaptive e-learning system . agent in the system, which in charge of investigating the learner who first use the system, providing a final test while the learner finishes the whole learning process and storing these resulting. The emerging e-learning is reshaping the instructional community and provides tremendous cost savings for both instructors and learners[1,2,3]. But it lacks in interaction aspect, intelligence,. for Adaptive e-Learning Wanjie Liang, Jianmin Zhao, and Xinzhong Zhu College of Mathematics, Physics and Information Engineering, Zhejiang Normal University, Jinhua, China wanjie.liang@zjnu.net,