A generic architecture for intelligent simulation training systems

Abstract

A generic object-oriented model for building intelligent simulation training systems with adaptive, interactive, pseudo-real-time simulations is presented. The model provides for the generic modeling of simulation scenarios that train the student to detect and remove undesirable situations, and to create desirable situations. These situations are organized, created (as a tutorial activity), detected, traced, supported and explained by a lattice of generically defined curriculum objects. Simulation objects are modified by the student whose actions may remove or delete situations. The feedback of the system and the evaluation procedures are based on the timing of the student actions, their side effects and their comparison with expert action suggestions. Activated segments of a distinct rule-based domain-expert system generate suggestions and evaluations. >