Distributed Intelligent Tutoring System Architectures

Abstract

Traditional architecture of Intelligent Tutoring Systems (ITSs) does not offer sufficient modularity. There is a lack of open distributed ITS architectures, despite ITSs being systems that may need frequent changes due to the modifications of particular course or adaptation to new courses. The chapter focuses on usage of distributed technologies in development of open ITSs to increase the modularity of the ITSs and facilitate the implementation of needed changes into ITSs. The aim of the chapter is to propose open and highly modular ITS architectures, using two distributed paradigms – intelligent software agents and services. To realize intelligent tutoring various types of learning materials and problems have to be presented to the learner, moreover it should be done intelligently enough to successfully simulate the human tutor. Thus all known ITSs concentrate on a certain problem domain or course to provide specific functionality for problems and examples of the domain. For example, the system has to be capable to analyse learner’s actions during the problem solving. Each new type of problems needs corresponding code to handle it. Problems differ from course to course and may change if the course is changed. As a consequence the functionality of ITS may be modified to adapt to the changes in the course or to a new course. The system should be open for certain types of components handling new types of problems, materials, feedback, etc. The architecture of ITS should support such an openness. ITSs traditionally have modular architecture consisting of four modules, namely tutoring module, expert module, student diagnosis module and communication module (Grundspenkis & Anohina, 2005). The main principle is to build components using only one type of knowledge (pedagogical knowledge, domain knowledge and knowledge about the learner). As a result, the architecture does not have sufficient modularity for complex ITSs. To facilitate modularity and change management distributed technologies like services and agents are used in ITSs. Well known examples of agent based ITSs are Ines system for nurse education (Hospers et al., 2003), Formal Languages and aUTomata Education system FLUTE (Devedzic et al., 2000), IVET virtual training environment (de Antonio et al, 2005) and WADIES – a Weband agent-based adaptive learning environment for teaching compilers (Georguli et al., 2003). Grundspenkis and Anohina (2005) have concluded that agent based ITSs mainly implement traditional modules as sets of agents. The authors have defined customizable set of agents that implements the system as a set of distributed components and at the same time maintains the traditional