Efficient evaluation of location-dependent skyline queries using non-dominance scopes

Abstract

With respect to a given query point q, a location-dependent skyline query (LDSQ) finds all objects that are not spatially dominated by any other in a specified set of objects. An object o is considered to spatially dominate another o if o is not worse than o for all search criteria and strictly better than o for at least one of the search criteria. Here, the search criteria include the objects' nearness to q and their non-spatial attributes. LDSQs have many valuable applications that need to find the best and nearest objects. Nevertheless, efficient evaluation of LDSQs faces challenges due to exhaustive comparisons among objects in determining their spatial dominance relationships which in turn depend on query points at the run time. In this paper, we introduce a notion of non-dominance scopes. For each object, a non-dominance scope is an area in which query points find the object not spatially dominated by other objects. With non-dominance scopes, LDSQs are simplified as searches for objects whose non-dominance scopes enclose query points. We develop algorithms to compute non-dominance scopes and evaluate LDSQs using indexed non-dominance scopes. We also evaluate our proposed approach. The experiment results show the outstanding performance of our proposed approach.

Efficient Index-Based Approaches for Skyline Queries in Location-Based Applications

“…Location-dependent skyline query (SQ). Given a query point q and an object set O, an SQ (which has also been studied in [10], [11], [14], [17], [27], [29]) finds all objects that are not locationally dominated by others with respect to q, formally, SQðq; OÞ ¼ fo j o 2 O; 6 9 o 0 2OÀfog o 0 ' q og. In the running example, an SQ issued at q 1 returns fo 3 ; o 4 ; o 5 g. .…”

“…Since locational dominance relationship among objects with respect to all possible locations in S is precomputed, object comparisons for dominance checks are completely waived at search time. The notion of nondominance scope was first introduced in our previous work [14] to answer location-dependent skyline queries. Here, we extend the idea to support other queries and explore parallel processing to boost search efficiency.…”

“…Thus, many recent research works, for example, [10], [11], [14], [17], [27], [29] have explored the notion of locational dominance in location-dependent information access.…”