BrowserUk:
Nothing, really. It just mentions them.
If the dataset were a set of straight-line distances, then it could warn you that the data has an error in it. My code doesn't really care, though. If the triangle is "bad", I'll still calculate the intersection point where the lines *would* meet, were they long enough.
...roboticus
When your only tool is a hammer, all problems look like your thumb.
| [reply] |
Nothing, really. It just mentions them.... I'll still calculate the intersection point where the lines *would* meet, were they long enough.
Oh, okay.
Though I have to say that it sounds like you are now performing 3 "triangle inequality" checks for every possible triangle (for my 17 node dataset that is 3*15! or nearly 4 Trillion additions and 4 Trillion comparisons) for no real purpose. Ie pure make-work....
For Dirk80's 24 node dataset, that rises to 1.8 trillion trillion redundant calculations.
With the rise and rise of 'Social' network sites: 'Computers are making people easier to use everyday'
Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
"Science is about questioning the status quo. Questioning authority".
In the absence of evidence, opinion is indistinguishable from prejudice.
| [reply] |
BrowserUk:
We're only picking three points for each triangle, and the point order doesn't matter, so the number of calls to chk_triangle_inequality during the exhaustive check for your dataset is n!/((n-3)!*3!). For your dataset, it's 680 calls, and 2024 for Dirk80's dataset. If it were trillions of checks, I'd've pressed ^C long before it ended. ;^D
Currently the check is pointless busywork, but since TSP problem is interesting to me, and printing interesting clues might spur ideas, so I threw it in for fun. I frequently litter my programs with print statements showing intermediate steps, or call functions whose results may be interesting. I delete it when it proves valueless to me, and then clean it up when I'm through. But I can't always predict beforehand which functions I'll find to be junk.
For example, when I was playing with your Bloom filter example, I put various print statements in my program, and noticed that one of the values was approaching an "interesting" value, allowing me to figure out the distributions.
Update: As AM said, wrong formula! Fixed.
...roboticus
When your only tool is a hammer, all problems look like your thumb.
| [reply] |