Priority R-tree

The Priority R-tree is a worst-case asymptotically optimal alternative to the spatial tree R-tree. It was first proposed by Arge, De Berg, Haverkort and Yi, K. in an article from 2004.[1] The prioritized R-tree is essentially a hybrid between a k-dimensional tree and a r-tree in that it defines a given object's N-dimensional bounding volume (called Minimum Bounding Rectangles - MBR) as a point in N-dimensions, represented by the ordered pair of the rectangles. The term prioritized arrives from the introduction of four priority-leaves that represents the most extreme values of each dimensions, included in every branch of the tree. Before answering a window-query by traversing the sub-branches, the prioritized R-tree first checks for overlap in its priority nodes. The sub-branches are traversed (and constructed) by checking whether the least value of the first dimension of the query is above the value of the sub-branches. This gives access to a quick indexation by the value of the first dimension of the bounding box.

Performance

Arge et al. writes that the priority tree always answers window-queries with \, O((N / B)^{1-1/d} + T / B) I/Os, where N is the number of d-dimensional (hyper-) rectangles stored in the R-tree, B is the disk block size, and T is the output size.

Dimensions

In the case of N = 2 the rectangle is represented by \, ((x_{min}, y_{min}), (x_{max}, y_{max})) and the MBR thus four corners \, (x_{min}, y_{min}, x_{max}, y_{max}).

See also

References

  1. ↑ L. Arge; M. de Berg; K. Yi (2004). "The Priority R-Tree: A Practically Efficient and Worst-Case Optimal R-Tree" (PDF). SIGMOD. Retrieved 12 October 2011.