The functional architecture and interaction model of a generator of intelligent tutoring applications

Abstract

This article presents the functional architecture and the interaction model of GENITOR, a generator of Intelligent Tutoring Systems (ITSsj. The major design issues that arise during the functional design and development of an ITS-generator concern the provision of mechanisms capable of supporting the description of the teaching subject domain and the specification of pedagogical features of the ITSs that it produces. The system described in the article provides ready-to-use yet flexible solutions for the instructional features of the developed ITSs. In this way, authors can concentrate on the compilation, description, and design of training material that relates directly to their area of expertise. Interaction between GENITOR and authors has been designed using IDFG, an interaction model that supports the representation of multiple aspects of interaction, including data flow, control flow, and task decomposition. As a consequence of using this model, the tasks that make up the authoring process are visualized by a number of authoring tools that are grouped into three subsystems: the reusability subsystem, the authoring subsystem, and the execution subsystem. The functionality of these subsystems and the tools they contain are described, as well as interaction with the overall system. Finally, conclusions drawn from the development of two ITSs using the system are presented in order to validate the design. 0 1997 by Elsevier Science Inc.

A framework for developing intelligent tutoring systems incorporating reusability

“…Several other efforts are being made to develop ITS shells and reusable ITS components, although they are not discussed here due to space limitations. For more on these efforts, see [1,7,10,12].…”

A hybrid expert system as an embedded module in tutoring systems

“…GENITOR (Kameas and Pintelas, 1997) is an ITS generator, that provides authors with tools that may be used to completely represent the knowledge domain of a training application, while ready-to-use solutions are provided for most of the authoring activities that relate to the specification of instructional strategy. It produces stand-alone intelligent training applications in subjects that are not necessarily related with each other which attempt to transfer to the trainees two kinds of skills: procedural knowledge on how to apply a certain methodology and declarative knowledge as the corresponding theoretical background to support it.…”

The Agent Oriented Multi Flow Graphs Specification Model

“…It constitutes a modification and an extension of IMFG model (Kameas, 1995) both in its philosophy and syntax. Although IMFG is a formal model that has been originally proposed for interactive dialogue description or for the specification of interactive applications (Kameas & Pintelas, 1997;Zaharakis et al, 1995), it is modified and extended for reasoning about multi agent systems. Its well-defined formalism for task decomposition and its event handling mechanisms are adjusted for the description of an agent's goal decomposition and for plan representation mechanisms as well as for the communication of messages between agents.…”