Caching dynamic skyline queries

Abstract

Given a query tuple q, the dynamic skyline query retrieves the tuples that are not dynamically dominated by any other in the data set with respect to q. A tuple dynamically dominates another, w.r.t. q, if it has closer to q's values in all attributes, and has strictly closer to q's value in at least one. The dynamic skyline query can be treated as a standard skyline query, subject to the transformation of all tuples' values. In this work, we make the observation that results to past dynamic skyline queries can help reduce the computation cost for future queries. To this end, we propose a caching mechanism for dynamic skyline queries and devise a cache-aware algorithm. Our extensive experimental evaluation demonstrates the efficiency of this mechanism compared to standard techniques without caching. © 2008 Springer-Verlag.