Cognitive Artifacts for Intelligent Agents in MAS: Exploiting Relevant Information Residing  in Environments

   page       attach   
Emiliano Lorini

Besides using languages and direct communication, humans adopt various kind of artifacts as effective means to represent and share knowledge, and finally support knowledge-based cooperation in complex work environments. Similarly to the human case, we argue that an analogous concept can be effective also in the context of cognitive multi-agent systems (MAS). Based on previous work on artifact-based environment and A&A conceptual framework, in this paper we investigate the use of a special kind of artifacts, cognitive artifacts, as computational entities
designed to store, process and make available those information which is relevant for agents to coordinate their cooperative and distributed activities. After introducing the main concepts, we discuss some of the practical benefits of the approach through an experiment based on CARTAGO and Jason technologies, respectively a platform for developing artifact-based environments for MAS and for programming cognitive BDI-based agents, comparing different interaction styles for teams of goal-directed agents engaged in distributed cooperative works