Agent roles: From methodologies to infrastructures

Role is an important and useful concept to define the collaboration in agent systems, since it allows to abstract from physical agents and to focus on their behaviour. So, the role concept is exploited in agent methodologies and agents infrastructures, but often in different and discontinue ways. The aim of this paper is to map the role concepts proposed in different agent methodologies, and also mapping other concepts related to role. Moreover, we try to extend this mapping also to agent infrastructures, so to help developers choosing the right methodologies and infrastructures. Providing a comparison (and outlining the contrasts) will be important and useful as a starting point to understand and to try to unify foundational agent concepts.