Emerging technologies have made available very large data repositories, which may be unreliable for a given preference criteria. In order to be able to process these repositories, users may need to discard useless information based on some preference conditions. Different preference-based query languages have been defined to support the bases for discriminating poor quality data and to express user's preference criteria. In this paper, we consider the preference-based query language, Top-k Skyline, which combines the order-based and score-based paradigms. Thus, Top-k Skyline is able to identify the top-k objects w.r.t. a score function f among the ordering induced by a multicriteria function m. Several algorithms have been proposed to implement these two paradigms independently; however, the problem of efficiently evaluating Top-k Skyline queries remains open. In this work, we propose evaluation strategies for Top-k Skyline queries and we report initial experimental results that show the properties of our proposed solutions.
