| commit | 97d2c93a87fc1b351eec678e16eb8c2ba1edd600 | |
| author | Laurens Van Houtven <lvh@laurensvh.be> | |
| Thu, 9 Oct 2008 18:52:15 +0000 (9 20:52 +0200) | ||
| committer | Laurens Van Houtven <lvh@laurensvh.be> | |
| Thu, 9 Oct 2008 18:52:15 +0000 (9 20:52 +0200) | ||
| tree | a4c0f9b1c8e1a1bf245f692c0276a59fcaf9a1ea | treesnapshot (tar.gz zip) |
| parent | 4b7ad199c0eb3d6f46cbd8b7f0f4a13358b6ee35 | commitdiff |
Bugfix + new feature
You can now create random asymmetric distance matrices.
The enumerative solver always used 1 as a first vertex. This meant it
wasn't always entirely enumerative, and sometimes generated suboptimal
solutions. Currently the first vertex is also permuted. This means the
algorithm has one extra city to permute, and as such is slower -- this
is because the first one cut corners, not because this one is
inefficient.
On branch master
Changes to be committed:
modified: generateRandomCoordMatrix.m
new file: generateRandomDistanceMatrix.m
modified: ltspsEnumerativeSolver.m
You can now create random asymmetric distance matrices.
The enumerative solver always used 1 as a first vertex. This meant it
wasn't always entirely enumerative, and sometimes generated suboptimal
solutions. Currently the first vertex is also permuted. This means the
algorithm has one extra city to permute, and as such is slower -- this
is because the first one cut corners, not because this one is
inefficient.
On branch master
Changes to be committed:
modified: generateRandomCoordMatrix.m
new file: generateRandomDistanceMatrix.m
modified: ltspsEnumerativeSolver.m
| generateRandomCoordMatrix.m | diffblobblamehistory | |
| generateRandomDistanceMatrix.m | [new file with mode: 0644] | blob |
| ltspsEnumerativeSolver.m | diffblobblamehistory |