Abstract

This chapter proposes an architecture for the development of intelligent virtual environments for training (IVETs) which is based on a collection of cooperative software agents. The first level of the architecture is defined as an extension of the classical intelligent tutoring system architecture that adds a new world module. Several software agents are then identified within each module. They communicate among them directly via messages and indirectly via a common data structure that is used for the collaborative development of plans. Some details are provided about the most remarkable interactions that will be established among agents during the system’s execution. The proposed architecture, and its realization in a platform of generic and configurable agents, will facilitate the design and implementation of new IVETs, maximizing the reuse of existing components and the extensibility of the system to add new functionalities.