Interactions between subparts of a system have been recognized as the source of complexity in many fields ranging from physics, sociology, neurology, to software engineering. Agent-Oriented Software Engineering (AOSE) was born under the promise of conquering complexity and enabling the development of more complex software. However, current AOSE approaches do not provide enough engineering tools to deal with the complexity derived from interactions. More mature fields such as economy or component-based software systems have recognized that interactions present a predominant role in the determination of the desired outcome providing mature background that can be applied to AOSE. AOSE may improve its ability to deal with complex systems by improving the tools applied to manage agent?s interactions in the overall design of the system. In this paper, we justify this assessment and propose some principles to improve AOSE methodologies regarding complexity
展开▼
