Affiliations: Department of Computer Science and Engineering, Guru Nanak Dev University, Amritsar, Punjab, India. E-mails: [email protected], [email protected]
Abstract: Artificial Agent Societies are analogous to human societies in which a collection of agents (analogous to humans) are residing or inhabiting a specific locality and are interacting with each other, for some solving common/individual purpose. Social networks, electronic markets and disaster management organizations can be viewed as such artificial (open) agent societies and hence can be best understood as computational societies. The members of such artificial agent societies are heterogeneous intelligent software agents, which are operating locally, cooperating and coordinating with each other in order to achieve goals of an agent society. These open agent societies have some kind of dynamics existing in them, in terms of dynamics of Agent Migration, Role-Assignment, Norm-Emergence, Security and Agent-Interaction. All these dynamic aspects are very closely interrelated to each other, as change in one reflects changes in others also. In this paper, we have presented a survey of all these dynamic aspects of an open agent society at its design-time and have tried to relate them to its member as well as non-member (external) agents. We have also related them, to the migration of a new agent in an open agent society, i.e., how role-assignment will be done for it, various issues related to security and trust of a society upon its arrival, issues related to norm-identification and norm-emergencefor it and its communication with the rest of the member agents of the society, that also, during and after its migration. In addition to this, we have also outlined various research challenges and directions available in the area of dynamics of artificial agent societies.
