Agent Communication Languages for Information-Centric Agent Communities

As the complexity and application scope of agent-based systems increases, the requirements placed on Agent Communication Languages have also increased with it. Currently-available ACLs focus on agent-based systems in the domain of knowledge agents. Therefore, they lack certain facilities required to implement large, complex, and robust information-centric agent systems. These facilities are required for efficient information transfer, use of multimedia information, and data security. Furthermore, information-gathering agents tend to execute longrunning and/or complex tasks, and require ACL support for managing tasks and conversations. Lastly, information agents (and other types of agents) may require more flexible agent-level management and control, specifically in the areas of mobile resources and/or intermittentlyconnected users. In this paper we attempt to discuss the shortcomings of existing ACLs and provide specific solutions to address them. The proposed ACL architecture is based on issues that we have encountered with the InfoSleuth system at MCC and other related work.