A forced transplant algorithm for dynamic R-tree implementation

Abstract

Spatial access methods play a crucial role in spatial database management and manipulation. The R-tree and its variations have been widely accepted as some of the most efficient spatial indexing structures in recent years. However, neither considers storage utilization and the global optimization of a R-tree structure. Presented in this paper is a new optimization technique named forced transplant algorithm, which can improve the node storage utilization and optimize the R-tree overall structures at the same time. Our experiments show that the R-tree with our new optimization technique achieves almost 100% storage utilization and excellent query performance for all types of data. © Springer-Verlag Berlin Heidelberg 2006.