As the availability of spatial information increases and location-aware services and applications proliferate, the efficient processing of spatial queries is becoming more important. Keyword- or content-based querying introduces the need to weight results by relevance. Mobility and communications constraints add even more factors. This paper combines location, keyword-based relevancy, and mobility to address the problem of efficient processing of moving top-k spatial weighted-distance queries. The authors describe algorithms for computing safe zones for such queries, propose optimizations, and analyze them from the perspectives of computation and communication costs.

The algorithms are based on three main ideas: multiplicatively weighted Voronoi diagrams, the existence of a relevance metric for keyword queries on the data objects, and an index structure that supports efficient retrieval based on spatial proximity and relevance to queries (the IR-tree). Techniques for the efficient update of the result set are also described. The algorithms are evaluated in experiments that measure processing and communication performance with real point-of-interest datasets and simulated movement.

The problem studied in this paper is complex, but important for current and likely future applications pertaining to spatiotemporal information. The analysis in this paper is necessarily simplified in many ways, for example, by being restricted to point locations and an Euclidean distance metric. Other constraints and data distributions are likely to suggest additional optimizations or different techniques. Nevertheless, this paper reports in detail on some substantial and well-thought-out research, which should be worth the attention of academics and advanced developers interested in spatiotemporal information.