Normative Multi-Agent Enriched Data Mining to Support E-Citizens

Abstract

The current trends in the development of multi-agent systems indicate the possibility to apply the concept of multi-agent systems employing ontologies for encoding the systems domain knowledge and procedural knowledge. Within such structures, the use of knowledge discovery models may represent an enhancement to the systems functionality in the context of discovering relations that can support the users activities. Distributed data mining (DDM) concepts ([19, 11, 18]) demonstrate that multi-agent systems are capable of using knowledge discovery processes in a variety of ways in the context of the process being supported by the agent as well as in order to expand its knowledge. If this is the case, the knowledge discovery becomes an intrinsic component of the agents learning process. The application of norms to support the work of agents, associated with the idea of normative multi-agent systems ([30, 2, 5, 4]), may significantly boost the performance of such systems by directing the agents actions and determining the desirable states of the agent itself as well as of the group it is part of. The chapter aims to discuss the key aspects of the development of multi-agent systems and knowledge discovery systems, and to present a proposal for an architecture of multi-agent systems supported by knowledge discovery systems.