GTree: An Efficient Grid-Based Index for Moving Objects

Abstract

In mobile environments, tracking the changing positions of moving objects efficiently could substantially improve the quality of the Location Based Services. There arises the high demand for the indexes to support frequent updates. In this paper, we propose a novel grid-based index for moving objects, namely the GTree. Based on the recursive partition of the space and lazy maintenance, the GTree could maximize the stability of the index structure while minimizing the dynamic adjustment, and therefore significantly reduce the update overhead. Different from traditional top-down or bottom-up updates, we present a median-down approach, which could effectively reduce the number of disk access. As an alternative, a bulk-loading technique is introduced. The experiments show that the GTree has good update performance as well as query performance.