Modeling & understanding environment using semantic agents

Understanding environment is a very complex task. Modeling and development of components of the environment in a symbolic way permit to simplify and clarify functional parts of components and interactions. In order to have a complete system specification, a rigorous behavior description is needed. Different levels of behavior abstraction are taking into account. The objective of our semantic modeling is to enhance architectural design and reduce complexity. It permits to agents to understand the environment, manage events and adapt the architecture. All the concepts of the environment like the component models and their behavior models are stored under event frames written in knowledge representation language. We present in this document, a generalized meta-model of behavioral aspects, that indexes the various environment behaviors in three ontologies. We have fully linked abstraction level with modeling and execution of scenarios. We show how software semantic agents can be modeled to build any interactive architecture.