Integrating Objective & Subjective Coordination in FIPA: A Roadmap to TuCSoN
Andrea Omicini, Alessandro Ricci, Giovanni Rimassa, Mirko Viroli · 2003
Abstract— Subjective and Objective coordination can be inte-grated and exploited fruitfully in the same context. In this paperwe investigate such integration in the context of FIPA agents, aim-ing at exploiting the coordination services provided by TuCSoNcoordination infrastructure. I. T HE B ABEL ’ S T OWER OF C OORDINATION Research on coordination possibly represents the mostun-coordinated activity in the history of computer science.Roughly speaking, coordination research has developed alongtwo basically separated lines in the DAI and SE field, respec-tively. In the former, coordination was interpreted as an indi-vidual, psychological activity, performed by a component (typ-ically, an agent of a multi-agent system – MAS henceforth)trying to achieve its own subjective goals in the context of amulti-component system. In the latter, coordination was basi-cally regarded as normative activity performed by some part ofa multi-component system on behalf of the system’s designer– typically, by a coordination medium provided by an infras-tructure. With respect to components, two different views wereadopted on coordination: the first, where components are thecoordinating entities, the second where components are the co-ordinated entities.Not surprisingly, the first approach seems to better suit sys-tems whose components exhibit a high degree of autonomy (in-telligent agents being the most obvious example), whereas thesecond fits well application scenarios involving a finer compo-nent granularity (as typical in the case of mobile agents). Du-ally, scientific confusion was not the only problem producedby these divergent efforts. The SE approach (see for instanceM