Structured Social Agents

Five essential capabilities of intelligent social agents are analysed. It is stipulated, that these abilities are crucial for the cooperativity of an agent society and it is argued that a maximum of cooperativity can be reached in the society in case all five features exist at the same time. From these five basic abilities we derive the model of a structured social agent as used in the multi-agent testbed RATMAN. The modeling of a society of intelligent agents is realized in RATMAN by specifying different knowledge bases (KBs), one for each agent. The KBs not only contain the agents’ world knowledge, but they take some of the fundamental mental capacities of the agents into account, like the degree of communication facilities, learning skills or self-reflection. We describe a prototypical model of such a KB, whose essential structure is derived from the five basic abilities an intelligent social agent is alledged to have. The different knowledge packages are structured hierarchically from simple propositional knowledge to highly abstract (meta-) knowledge. They are realized with various representation formalisms (i.e. different kinds of logics), which are mutually compatible.