Student Record Prototype System

[Name of the Institute]

[Name of the Supervisor]

TABLE OF CONTENT

CHAPTER 4: DISCUSSION AND ANALYSIS3

CHAPTER 5: STUDY TWO: STUDENT PROBLEM-SOLVING BEHAVIOR8

Comparison15

Results of rater classification19

Training examples23

CHAPTER 6: CONCLUSION25

Future plans26

REFERENCES29

CHAPTER 4: DISCUSSION AND ANALYSIS

The area of artificial intelligence and knowledge-based systems offers great potential for enhancing the support of simulation modeling. Conventional simulation languages are limited by the necessity to settle on fixed, simplistic resolutions to a number of complex tradeoff decisions. The introduction of intelligence into such languages, however, promises to yield unprecedented levels of modeling support, providing a means to capture complex concepts and to adapt to different situations. Rather than proposing any new simulation language constructs, the research in this paper aims at providing an intelligent simulation modeling environment. To explore the potential of intelligent modeling support and provide a basis for future development, a project was designed to demonstrate that a knowledge-based simulation system could be developed in practical form.

Hayes-Roth, Waterman, and Lenat (1983) stress the importance of prototyping when explor-ing new areas of application for intelligent systems. Such research provides an important complement to theoretical and conceptual development by testing the practical feasibility of new concepts and suggesting extensions and identifying new research issues based on actual experience. To our knowledge, this is the first report of a routinely used expert system to support simulation modeling in a general modeling environment. An undergraduate course that emphasizes simulation modeling provided an ideal opportunity to build and test a prototypical system.

The large number of students (approximately 125 per semester) requires a high level of support, particularly in correcting conceptual errors in their models. However, this support is difficult to provide through the use of graduate student teaching assistants, who themselves possess various levels of modeling expertise. The consequent need to capture and provide modeling expertise was identified as a problem which was well-suited to the general goals of the research.

The exploratory project, therefore, focused on the design and implementation of a system which could support the students in debugging models. Although the knowledge-based system was developed in an instructional setting, the situation is, in fact, quite similar to that faced by a simulation practitioner. The practitioner may require the resolution of significant logical errors within a simulation model. Without expert assistance, he must resolve the problem within his own resources, perhaps seeking help from a colleague, vendor, or consultant. In any case, in practice the resolution of logical errors can be extremely expensive, both in the amount of time needed to resolve them and the potentially negative impact they can have on a simulation project.

The students' behaviors parallel those of practitioners. When faced with logical errors, they must also resolve their errors within their own resources and, when those attempts fail, seek help from teaching assistants and the course instructor. We expect that because the students are currently taking the course and the material is fresh in their minds, their errors will be even more subtle and difficult to diagnose than will a &dquo;typical&dquo; simulation ...