Many users and applications require the integration of semi-structured data from autonomous, heterogeneous Web sources. Over the last years mediator systems have emerged that use domain knowledge to overcome the problem of structural heterogeneity. However, many users of these systems do not have a thorough knowledge of the complex global schemas and of the comprehensive query languages. Consequently, easy-to-use query interfaces like keyword search and browsing have to be supported. The aim of the proposed PhD project is the index-based realization of keyword searches in concept-based mediator systems. In order to avoid unnecessary source queries an index structure is maintained on the global level and used during query planning and processing. 1 Research Problem Nowadays, many search requests require the usage of information from different, distributed, and autonomous (Web) sources. Applications can be found in scientific as well as e-commerce environments. Prominent examples are th...