Cognitive Agents As A Design Metaphor In Environmental-Knowledge Management

Representing and sharing knowledge has been a central problem in artificial intelligence since its inception. Representations such as semantic networks, frames, concept maps and ontologies, as well as various methodologies for using these systems have been proposed for dealing with such issues. However, problems exist about issues such as communication among heterogeneous agents, incomplete or uncertain knowledge, imprecise formalizations, and so on. Here, a mapping system between knowledge representations (concept maps and ontologies) is modeled using a methodology for the development of multi agent systems. Ontologies are formalized from non-formal concept maps and can then be used to represent agents knowledge, and to facilitate the communication among persons and software agents. A system is presented, in which a set of agents, implementing three functionalities (retrieval, disambiguation and formalization), collaborates in the process of knowledge management. This multi-agent system is part of a larger knowledge management system based on concept maps, and facilitates the use of generated and managed knowledge by not only people but also other software agents, namely those which require knowledge about domains that have been represented as concept maps, such as the environment, domain, object of this book.