A Formal Description Language for Multi-Agent Architectures
Multi-Agent Systems (MAS) constitute a highly promising software architectural approach for modern application domains such as peer-to-peer and ubiquitous computing, information retrieval, semantic web services or e-business. Unfortunately, despite consid
- PDF / 1,074,569 Bytes
- 21 Pages / 430 x 660 pts Page_size
- 117 Downloads / 191 Views
Information Management Research Unit, University of Namur, 5000 Namur, Belgium [email protected] 2 ISYS- Information Systems Research Unit, IAG-Louvain School of Management, Université catholique de Louvain, 1348 Louvain-la-Neuve, Belgium {kolp,wautelet,achbany}@isys.ucl.ac.be
Abstract. Multi-Agent Systems (MAS) constitute a highly promising software architectural approach for modern application domains such as peer-to-peer and ubiquitous computing, information retrieval, semantic web services or ebusiness. Unfortunately, despite considerable work in software architecture during the last decade, few research efforts have aimed at truly defining languages for designing such architectures. This paper identifies the foundations for an architectural description language (ADL) to specify multi-agent system architectures. We propose a set of system design concepts based on the BDI (beliefdesire-intention) agent model and existing classical ADLs. We conceptualize it with the Z specification language to capture a “core” model of structural and behavioural elements fundamental to an architecture description for BDI-MAS. We partially apply it on a data integration system example to illustrate our proposal.
1 Introduction In the last few years, software applications have increased in complexity and stakeholders’ expectations principally due to new Internet-centric application areas. These areas demand robust software that can operate within a wide range of environments and can evolve over time to cope with changing requirements. Moreover, such software has to be highly customizable to meet the needs of different kinds of users and sufficiently secure to protect personal data and other assets on behalf of its stakeholders. Not surprisingly, researchers are looking for new software paradigms that cope with such requirements. One source of ideas that is gaining popularity for designing such software is the area of multi-agent systems (MAS) architectures. They appear to be more flexible, modular and robust than traditional architecture including objectoriented ones. They tend to be open and dynamic in the sense that they exist in a changing organizational and operational environment where new components can be added, modified or removed at any time. Research in this area has notably emphasized that an MAS architecture is conceived as a society of autonomous, collaborative, and goal-driven software components (agents), much like social organizations. M. Kolp et al. (Eds.): AOIS 2006, LNAI 4898, pp. 143–163, 2008. © Springer-Verlag Berlin Heidelberg 2008
144
S. Faulkner et al.
Such architectures become rapidly complicated due to the ever-increasing complexity of these new business domains and their human or organizational actors. Practitioners have come to realize that getting a complex architecture right is a critical success factor for the system life-cycle. They have recognized the value of making explicit architectural descriptions and choices in the development of new software: a rigorous architectural design
Data Loading...
