DYNASTAT: A Methodology for Dynamic and Static Modeling of Multi-agent Systems

Abstract

Multi-agent systems are increasingly being used within various knowledge domains. The need for modeling of the multi-agent systems in a systematic and effective way is becoming more evident. In this chapter, we present the DYNASTAT methodology. This methodology involves a conceptual overview of multi-agent systems, a selection of specific agent characteristics to model, and a discussion of what has to be modeled for each of these agent characteristics. DYNASTAT is independent of any particular modeling language but provides a framework that can be used to realize a particular language in the context of a real-world example. UML 2.2 was chosen as the modeling language to implement the DYNASTAT methodology and this was illustrated using examples from the medical domain. Several UML 2.2 diagrams were selected including a use case, composite structure, sequence and activity diagram to model a multi-agent system able to assist botha medical researcher and a primary care physician. UML 2.2 provides a framework for effective modeling of agent-based systems in a standardized way which this chapter endeavors to demonstrate.