| There is a growing demand for principled and useful instructional software applications in both academic and industrial settings. The need for effective tutoring and training is increasingly important in technical fields, which demand the learning of complex tasks and the use of large knowledge stores. In the last two decades, intelligent tutoring systems (ITSs) have been proven to be highly effective as learning aides, and numerous ITS research and development efforts were initiated. However, few tutoring systems have made the transition to the commercial market. The main reasons for this failure to deliver are that the development of ITSs is difficult, time-consuming, and costly. There is a need for easier, more cost-effective means of developing tutoring systems.; In this dissertation, a novel methodology and architecture for generating intelligent tutoring systems for a wide range of domains is described. The architecture incorporates an ITS shell that interacts with any generic task-based expert system to produce a tutoring system for the domain knowledge represented in that system. The focus is on the issue of reusability. The knowledge-rich structure of generic tasks can be reused for instructional purposes, allowing the tutoring of domain and problem solving knowledge embedded within an expert system. The integration of this reusable knowledge with other reusable ITS components creates a powerful environment for the generation of tutoring systems for different domains.; The architecture developed has many features needed in an effective ITS authoring environment. Among other features, it employs a runnable deep model of domain expertise, facilitates fine-grained student diagnosis, and offers an easy method for building expert systems and generating ITSs from them. It was used to generate two tutoring systems: an Excel tools tutor and a composite materials fabrication technology tutor. Evaluation studies of the architecture and tutoring systems generated showed that the architecture allows ITS authors to generate instructionally effective tutors in a simple and straightforward process. The central contributions of this research are: (1) an architecture that generates intelligent tutoring systems from generic task expert systems, (2) a proof of concept implementation of the architecture, and (3) a concrete demonstration of the reusability of the knowledge stored within a generic task knowledge-based system. |
