Framework for Acquisition, Representation, and Use of Knowledge, Learning, and Collaboration among Intelligent Multi-Agent Systems in the Domain of Internet Infrastructure (SKAU)

This article explores the definition and evolution of Computational Agents (CAs), from their basic conceptualization as entities that interact with the environment through sensors and actuators, to their sophistication with the use of Artificial Intelligence tools. The article highlights that despite their increasing autonomy, CAs lack human intelligence and emphasizes the ethical responsibility of programmers in their development. Furthermore, the work proposes a specific architecture for multi-agent systems, the A2RD (Autonomous Architecture over Restricted Domains), detailing its layers and operation. The article also discusses the application of CAs in Internet infrastructure, specifically in routing information management, through the SKAU environment, which facilitates collaboration among agents and the construction of a specialized knowledge base. Additionally, the article addresses challenges related to the identification, authentication, and accountability of agents in complex systems and points to directions for future research, including the development of collaboration protocols, ethical evaluation of agents, integration with language models, and practical applications in various domains.