Ontological Basis for Agent ADL Stéphane Faulkner and Manuel Kolp 1 Information Systems Unit, Université catholique de Louvain, 1, Place des Doyens, 1348 Louvain-La-Neuve, Belgium {faulkner, kolp}@isys.ucl.ac.be Abstract. Multi-Agent Systems (MAS) architectures are gaining popular- ity over traditional ones for building open, distributed or evolving soft- ware. To formally define system architecture and reasoning about it, nu- merous architecture description languages (ADLs) have been proposed during the last decade. However, few research efforts aim at truly defining a description language for MAS architectures. The paper introduces an ontological basis aimed at capturing a “core” set of structural and behav- ioral concepts, and their relationships, we consider fundamental to define an ADL for MAS architectures. 1 Motivation The explosive growth of application areas such as electronic commerce, knowledge management, peer-to-peer and mobile computing has profoundly changed our views on information systems engineering. Systems must now be based on open architec- tures that continuously evolve to accommodate new components and meet new re- quirements. These new requirements call, in turn, for new concepts and techniques for engineering and managing information systems. For these reasons – and more – Multi-Agent Systems (MAS) architectures are gaining popularity over traditional systems, including object-oriented ones [5]. To cope with the ever-increasing complexity of such architectures, there have been, through the last decade, different proposals for providing a sound basis for formally describing system architecture and reasoning about it. In particular, a num- ber of Architecture Description Languages (ADLs) have been proposed (see e.g., [2]). ADLs provide constructs for specifying architectural abstractions in a descriptive notation. They offer formal mechanisms for decomposing a system into architectural elements, specifying how these elements are combined to form a configuration. Unfortunately, few research efforts have aimed at truly defining an ADL for MAS architectures. Therefore, we introduce here an ontological basis that take the Belief- Desire-Intention (BDI) agent model into account to identify ADL concepts for speci- fying MAS architectures. 173