cvt_test

cvt_test, a MATLAB code which calls cvt(), which creates Centroidal Voronoi Tessellation (CVT) datasets.

Licensing:

The information on this web page is distributed under the MIT license.

Related Data and Programs:

cvt, a MATLAB code which computes elements of a Centroidal Voronoi Tessellation (CVT).

Source Code:

• cvt_test.m, calls all the tests.
• cvt_test.sh, runs all the tests.
• cvt_test.txt, the output file.

• cvt_test01.m works in 2D, computes 10 points, using UNIFORM initialization and UNIFORM sampling, IT_FIXED is 1, 10000 sample points are used, and the random number seed is 123456789.
• cvt_test02.m repeats test 1 with twice as many iterations.
• cvt_test03.m repeats test 1 with 100 times as many sample points.
• cvt_test04.m repeats test 1 with UNIFORM initialization and HALTON sampling.
• cvt_test05.m repeats test 1 with UNIFORM initialization and GRID sampling.
• cvt_test06.m repeats test 1 with UNIFORM initialization and RANDOM sampling.
• cvt_test07.m repeats test 1 with a different seed.
• cvt_test08.m repeats test 1 with a different batch size.
• cvt_test09.m repeats test 1 with IT_FIXED = IT_MAX.
• cvt_test10.m generates 100 points in 3D.
• cvt_test11.m experiments with a Cartesian grid as starting point.
• cvt_test12.m tests the 'RANDOM' initialization option.
• cvt_test13.m works in 2D, computes 100 points, using USER initialization and USER sampling, IT_FIXED is 1, 10000 sample points are used, and the random number seed is 123456789. The user geometry is the unit circle, with uniform density.
• cvt_test14.m works in 1D (where we know the exact answer) and shows that it can take a long time for the computation to get close to the exact answer.
• cvt_circle.txt, a set of 100 CVT points in a circle, generated using the built in USER routine.
• cvt_circle.png, a PNG image of the 100 CVT points in a circle.