| | |
Summary: Blocked Randomized Incremental Constructions
Nina Amenta and Sunghee Choi
Technical Report number TR-02-54
University of Texas at Austin
Abstract
Randomized incremental constructions are
widely used in computational geometry,
but they perform very badly on large data
because of their inherently random mem-
ory access patterns. We define an inser-
tion order which removes enough random-
ness to significantly improve performance,
but leaves enough randomness so that the
algorithms remain theoretically optimal.
1 Introduction
A look at recent textbooks [1, 2] shows that
randomized incremental algorithms are a
central part of computational geometry.
Many randomized incremental algorithms
construct geometric structures; one of par-
|
