Using intelligent agents to combine heterogeneous distributed data

Effective interaction with distributed databases means providing the user with location transparency, i.e. it should appear as if the data is located at a single site. One major challenge in obtaining location transparency is the integration and presentation of heterogeneous, distributed data in near-real time with a user-friendly Web-based interface. This paper describes two projects that apply intelligent agents to the problem of integrating and presenting heterogeneous distributed data. The first project is a Web-based, simulation scenario-generating application that integrates data from distributed sources to build a simulation scenario for an interactive wargaming simulation called Janus. This scenario-generating application is called the Dynamic Scenario Builder (DSB) and will ultimately integrate data from the US Army's Functional Description of the Battlespace (FDB) with Janus or other simulation databases to generate simulation scenarios. The second project involves the development of a system to enable logistics planners to access distributed databases from remote locations and autonomously update data sources such that modifications are dynamically mirrored and displayed to all users of the distributed system. A key feature of this research is the visual presentation of the summarized information in a meaningful manner such that "drilling down" to more detailed information is evident. This project applies a platform for mediators called HERMES (Heterogeneous Reasoning and Mediator System) as a means of data integration from several heterogeneous distributed databases.