Specifying Multiagent Systems in E-Learning Environments

 < Day Day Up > 



Let us examine the multiagent system for course maintenance and recommendation that was designed in the literature (Lin et al., 2003). In this system, online course materials, including all textbooks or e-books and study guides, plus the project handouts, can be downloaded and installed on students’ hard drives. Students only need to be online to post material on the conference, do the quizzes, or send and receive e-mails pertaining to the course. The online course materials are updated often in order to keep them as current as possible, especially in some rapidly changing fields, like “computing and information systems.” Because of the complexity of the materials, and the short development cycles within which they are produced, the course instructor should make the necessary adjustments time to time for the benefit of the students. Whenever there is a significant change in the content of several designated Web pages of online course materials, students who take the course should be notified by the course coordinator by e-mail. Figure 6 shows the conversation schemata for course maintenance.

click to expand
Figure 6: A conversation schemata for course maintenance

The conversation model of the course material change notification consists of the following elements. For simplicity of illustration, we assume that a student who takes the course is in either of the three phases, numbered 1, 2, or 3. The interpretation of the phases is trivial and left undefined (for example, Phase 1 might be the phase before the first exam, Phase 2 the phase between the first exam and the second exam, and Phase 3 the phase between the second exam and the final exam), except that we assume only students who have passed the previous phase are allowed to enter the next phase. A course Web page also bears a phase number, indicating to which phase its content is significant. Once a change is made to a Web page, all students taking the course and who have a phase number that matches the phase number borne by the Web page will be sent the link to that page.

  • Notification Agent Control Client (NACC): The NACC of an instructor or a student runs on his or her machine and allows him or her to control the behavior of the corresponding Notification Agent deployed in a distributed environment. In our system, NACC adds a student into the student database or removes the student from the database or changes the phase number the student is currently in.

  • Notification Agent (NTFC): The basic function of the Notification Agent is to send e-mails to students taking the course, according to the student profiles stored in a database, when the course material has been significantly changed.

  • Monitoring Agent (MNTR): The Web Change Monitoring Agent of a system administrator monitors a collection of course material URLs stored in a database. When the agent detects a significant change, it sends a message to the Notification Agent. Also, once a broken link is detected in the topic tree, it notifies the maintenance agent to either correct the link or delete the orphaned page.

  • Student Information Agent (STIF): A Student Information Agent is designed for providing services about student information, such as providing an e-mail list for a course by automatically maintaining the e-mail list of students taking a course; and maintaining the profile of each student.

  • Maintenance Agent (MNTN): The maintenance agent provides proxy services to the instructor. It maintains the content of the topic tree.

  • Topic Tree or Link Database (LINK): The course material is organized in the form of a topic tree. Each entry in the topic tree is a link to a Web page.

Let INST and STUD denote the multisets of instructors and students, respectively, and I, S, and L denote the instructor (we assume that there is only one instructor), the initial roll of the class, and the initial content of the course (in the form of the set of links), respectively. Figure 7 shows the Gamma program that specifies the above system:

click to expand
Figure 7: The course material maintenance program

In this program, constants are written in boldface words. Each student record is a tuple (student, phase, mailbox), where student is the name of the student, phase the phase number where the student is in, and mailbox the mailbox of the student, which is a multiset of e-mail messages. Each entry of the link database is also a tuple (link, phase, status), where link is the link to the Web page in the topic tree, phase the phase number this page is designed for, and status the status of the page, which can be either normal, changed, or broken. Boolean functions Enroll(s) and Drop(s) return whether student s is enrolled in the class or wants to drop. Modified(l) function returns whether a particular Web page pointed to by link l has been modified or not. Pass(s, p) function finds out whether student s has passed phase p or not. Add(l, i) function indicates whether instructor i wants to add the page pointed to by link l into the link database or not. Change(l, i) function returns the link to the changed page, with its original pointed to by l. Update(l, i) function updates the broken link l and returns the corrected link.

The program consists of configurations in two levels: the MAIN configuration in the higher level and all other configurations in the lower level. Program P in MAIN configuration exchanges elements of the multisets in the environments of the lower-level configurations.

This example shows how the Gamma language expresses the architecture of a multiagent e-learning system succinctly. With the underlying computing model, we do not need to consider the specifications of nonessential features of the system, e.g., the number of program units, connection links for communications, and organizations of data, and therefore, can focus on the specification of the overall architecture. It catches the way program units interact with one another, and local computations, such as the implementations of those local functions, are left to the subsequent design phase.



 < Day Day Up > 



Designing Distributed Environments with Intelligent Software Agents
Designing Distributed Learning Environments with Intelligent Software Agents
ISBN: 1591405009
EAN: 2147483647
Year: 2003
Pages: 121

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net