Dynamic spatial approximation trees
Journal of Experimental Algorithmics (JEA)
Efficient parallelization of spatial approximation trees
ICCS'05 Proceedings of the 5th international conference on Computational Science - Volume Part I
| Hi-index | 0.00 |
The Dynamic Spatial Approximation Tree (dsa-tree) isa recently proposed data structure for searching in metricspaces. It has been shown that it compares favorablyagainst alternative data structures in spaces of high dimensionor queries with low selectivity. The dsa-tree supportsinsertion and deletions of elements. However, it has beennoted that deletions degrade the structure over time, so thestructure cannot be regarded as fully dynamic in the sensethat deletions are not sustainable for long periods of time.In this paper we propose and study a new method to handledeletions over the dsa-tree, which is shown to be superiorto the former in the sense that it does not affect searchtime at all. Indeed, we show that the resulting tree is exactlyas if the deleted element had never been inserted. Theoutcome is a fully dynamic data structure that can be managedthrough insertions and deletions over arbitrarily longperiods of time without any reorganization.