I just finished an implementation of the Delaunay's incremental flipping algorithm. This algorithm has a time complexity O(N log N)
.
The application of the algorithm is based on taking each point as an antenna of a telephone company. Using the Delaunay's algorithm, I have to triangulate the space with such points then generate a Voronoi diagram using the triangulation in which each Voronoi polygon represents the coverage of each antenna
Now, I must solve the following problem:
For each given point, and a constant
d
, relocate all the points in the plane, without exceeding the distance d from the original coordinates of each point, so that the Voronoi regions are maximized.
I can not imagine how to solve this problem with an efficient algorithm. Delaunay's and Voronoi's algorithms are already implemented.
Thanks!
You can try Lloyd's algorithm. For each site compute the center of gravity and compare it with your old site. If the distance doesn't exceed the constant d relocate the site otherwise do nothing. Retriangulate the sites to smooth the mesh.