ארנון שטורם

Agent-based systems have evolved during the last two decades. To support the development of such systems, agent-oriented methodologies have emerged. In general, most of the methodologies have originated from two major research domains, namely software engineering and artificial intelligence, and were adjusted to address the agent abstraction. It seems that many of the methodologies share a common basis, an observation that calls for unification and for standardization. In this chapter, we survey existing agent-oriented methodologies and describe the support for agent-based concepts required in such methodologies. We then analyze the most influential agent-oriented methodologies in light of the required agent-based concepts as well as other criteria. We also examine alternatives such as methodology integration and the support for developing a tailored agent-oriented methodology. The main concern that arises from the survey and the analysis is the lack of evaluation of agent-based methodologies, which may have negatively affected, at least in part, the adoption of these methodologies for developing agent-based systems. We also discuss the need to further extend the methodologies to support the entire lifecycle.