Probabilistic Methods for Data Integration in a Multi-Agent Query Answering System

This report describes a proposal for a multi agent ontology-mapping framework that makes use of probabilistic information in order to enhance the correctness of the mapping. The proposed research focuses on two correlated areas namely similarity measures with its representation as a Dempster-Shafer belief function and usability of different optimalisation methods for combining these belief functions in a distributed environment. The main goal of our proposed research is to establish a multi agent framework that integrates user query related information from distributed scientific databases utilizing the AQUA system. The outcome of the research will contribute to the feasibility study of a distributed information integration network that is based on the European Commission Joint Research Centers data management and dissemination databases (AlloysDB, GasketDB, CorrosionDB, HTR-FUELDB), which stores mechanical and physical properties of engineering materials produced by the European RTD projects. These databases cover the materials behavior at low, elevated and high temperatures for base materials and welded joints and also includes irradiation materials testing in the field of fusion and fission and thermal barrier coatings tests for gas turbines.