Applicability for Higher Dimensions



next up previous contents
Next: 7.1.5 A Practical GRS Up: 7.1.4 The KIRKPATRICK Point Previous: Efficiency Considerations

Applicability for Higher Dimensions

The principles of the KIRKPATRICK point location method can be applied to higher dimensional grids too. The EULER formula as a basis for the proof of the KIRKPATRICK method

(the sum of vertices minus edges plus faces is two, including the unbounded face) can be generalized under the assumption of a simply connected region (i.e. a region with no ``holes'' in it) as

where denotes the (geometric) dimension and the respective geometric primitive object of dimension (i.e. a point in zero-dimensional space, a line in one-dimensional space, a triangle in two-dimensional space, a tetrahedron in three-dimensional space, etc.). The proof for the KIRKPATRICK method can be conducted in the same way, and point location will still be of time. However, it is not clear if the SDAG can be constructed in time too, as it is the case in two-dimensional space. Especially techniques for retetrahedralization in linear time in higher dimensions can not obviously be carried out using similarly structured algorithms as in two dimensions [Kirk83].

For TCAD applications, the three-dimensional case is of special interest. Assuming an initial tetrahedral grid, point removal would result in a star-shaped solid bound by topological two-dimensional geometric primitives lying in three-dimensional space, i.e. triangles. This star-shaped solid is now retetrahedralizated, and the new tetrahedrons point to all old intersecting tetrahedrons. While in two-dimensional space removal of a vertex and retriangulation will remove a total of two triangles, point removal in three dimensions results in a total of three removed tetrahedrons. Therefore the yield is for a star-shaped solid initially filled by tetrahedrons.



Martin Stiftinger
Tue Nov 29 19:41:50 MET 1994