A self-organized multiagent approach for distributed management of contextual data

Pervasive computing devices are able to generate enormous amounts of data, from which knowledge about situations and facts occurring in the world should be inferred for the use of pervasive services. However accessing and managing effectively such a huge amount of distributed information is challenging for services. In this paper we propose a self-organized approach to autonomously organize distributed contextual data items into sorts of knowledge networks. Knowledge networks are conceived as an alive self-organized layer in charge of managing data, that can facilitate services in extracting useful information out of a large amount of distributed items. We motivate our approach and present the W4 Data Model that we used to represent contextual data. On these basis, we introduce the general idea of W4 Knowledge Networks and we detail the specific bio-inspired approach for self-organized networking of data items. A case study is introduced to clarify the concepts expressed, and experimental results are reported to support our arguments and proposal.