Project description

The success of Semantic Systems has led to large instantiations of ontologies which serve as a description of various systems and state of affairs. E.g. such instantiations (also called configurations) specify the structure and properties of Web sites, business processes or technical systems and may comprise more than one million entities. In this context, ontologies describe the allowed configurations and their costs.

However, ontologies are not stable but change over time. In this case, the configurations have to be changed in order to reflect the alterations of the ontology. This change is represented by a modification of the ontology instantiation. The generation of this modified instantiation is a complex optimization process for various reasons. In particular, this reconciliation process (called instance reconciliation) must decide which legacy instances can be reused and which new instances must be generated such that the overall set of instances obey the constraints of the new ontology. In addition, the costs of changing a configuration must be minimized.

Consequently, a major aim of RECONCILE is to develop a reconciliation language on top of a logical description language that allows the expression of change-knowledge and change-costs. Because of the size of the problems and our practical experiences regarding the expressivity of required language constructs, we expect as a major challenge for RECONCILE the development of feasible problem solving methods. Practical experiences in similar domains (like constraint-based configuration) show the high utility of heuristic search as well as problem decomposition techniques.

Consequently, we will explore the application of these techniques for the reconciliation of instances with changed ontologies. Moreover, we will support the users with debugging and repair methods for formulating change-knowledge. All methods are evaluated by proof-of-concept prototypes and examples of real world applications.

RECONCILE was awarded for the best proposal of FIT-IT Semantic Systems Call 2009.

This project is funded by the Austrian Federal Ministry for Transport, Innovation and technology (BMVIT) and by the Austrian Research Promotion Agency (FFG) in the programme "FIT-IT Semantic Systems" (grant number 825071).

For further details visit our project's website