School of Computer, National University of Defense Technology, Changsha, Hunan Province, China 410073

---

Abstract—Operating in highly dynamic and unpredictable environments with partial requirements, modern software systems are characterized by context-awareness, flexible organization and dynamic interactions. The increasing complexity desires natural high-level abstractions and effective programming mechanisms to facilitate the development and maintenance of such systems. However, the abstraction, composition and interaction mechanisms in current programming approaches do not directly support those characteristics. This paper proposes an organizationbased agent-oriented programming (OrgAOP) approach that treats organizations, roles and groups explicitly, and provides mechanisms to support context-awareness, flexible organization and dynamic interactions. This paper describes the philosophy of OrgAOP, designs a programming language OragentL with explicit language facilities supporting OrgAOP and illustrates how to program and implement MAS with OragentL by a case study.

Index Terms—Agent-oriented programming, Organization theory, Organization-based programming, Dynamics.

[PDF]

Cite: Cuiyun Hu, Xinjun Mao, Yin Chen, "Programming Dynamic and Open Multi-Agent Systems with Organization Metaphor," Journal of Software vol. 8, no. 1, pp. 151-159, 2013.