Agents and ontologies for a smart management of heterogeneous data: The IndianaMas system

The IndianaMas system is a platform for storing, retrieving and analyzing images, manual sketches and multilingual texts about rock carvings: to manage heterogeneous data formats, languages and data sources, it adopts a multiagent architecture that makes easier coping with such a mixed environment, and exploits ontologies to have a clear, formal and self contained representation of the domain. These two architectural solutions make possible to have a very modular organization of the overall system and to manage in parallel different kinds of data, which are exposed thanks to a digital library. In this paper we present the architecture and the functionalities of the IndianaMas system, focusing on the coordination level between the agents and on the organization of structured data in the digital library, to show how the adopted architectural approaches and solutions really allow the management of such a complex system.