This report describes the results of practical and theoretical investigations for developing representation systems which support the construction of declarative models, which can be checked for consistency and which can be used as the basis for problem solving processes in different application contexts. The report is structured into four parts, each with its own set of references in order to provide an overview on the state of the art for each of the areas covered. First, in Part 1, we describe research results about expressive description logics, then, in Part 2, optimization techniques for instance retrieval are discussed in the context of grounded conjunctive queries. Afterwards, in Part 3, the expressivity of the query language is enhanced in order to provide a combination of spatial and ontological reasoning. In the last part, we focus on a middleware architecture to provide efficient and scalable reasoning systems for large-scale applications.