Modeling And Services For Adaptive Collaborative Delivery Of Annotated Multimedia Resources

Adaptive systems are becoming essential for supporting the overload and the diversity of digital documents to be archived, retrieved and disseminated. They indeed represent the most promising solution for individuals to be able to handle the amount of data which they are daily flooded with during their professional activities or on their personal devices. However, providing adaptive management of heterogeneous resources remains an important research issue as it requires extensive and global environmental knowledge management to be effective. Communities, as they are sharing interests and accesses to resources, present very interesting characteristics that enable systems to deliver automated and personalized services; the scope of such processes being to take advantage of collaborative involvement in order to provide relevant knowledge management to users, to ensure the consistency of data manipulation, and to improve the distribution of resources within communities. This dissertation presents a collaborative information management framework dedicated to the personalized delivery of multimedia documents. We first propose an Information Modeling for Adaptive Management (IMAM) which is an innovative set of algebraic structures making it possible to categorize and manipulate any piece of environmental knowledge that is useful for an enhanced delivery of data. In addition, this modeling is the generic basis for the definition of operators and services in collaborative environments; these functions rely on contextual information for personalizing the retrieval and the distribution of multimedia documents. Then, after ensuring an efficient storage management of annotations in XML with a mapping to the Extended Binary Graph data structure, we define a powerful indexing strategy that makes the retrieval of annotations faster. Finally, we design adaptive services based on IMAM’s formal modeling. These services comply with complex distributive models structures such as Peer-to-Peer and perform personalized query optimization and data placement in collaborative environment.