Hi guys;
Miss mi; you were close
but don't give up the day job!
Rap:
After awhile I got this one
.......1
.5 .6 12 .3
..10.. 11
.9.. 7.. 8. 2
.....4
But it isn't unique
The method is finding all permutations of four integers less than 13 and greater than 0 that sum to 26, and then setting these permutations into the star of david pattern to find the ones that fit. Since there are 32 permutations there are more than one pattern.
Cool!
TTH:
........1
9.10..2.5
....7.....12
..8..4..3..11
........6
I never doubted you for a second!
"
now I know how to show what I got."
Oh boy! Let me get a front row seat and popcorn before you let them puppies out!
Starting from the top clockwise:
12, 6, 9, 5, 3, 11, 1, 10, 2, 8, 7, 4.
Which gives:
12, 6, 5, 3.
9, 5, 11, 1.
3, 11, 10, 2.
1, 10, 8, 7.
2, 8, 4, 12.
7, 4, 6, 9.
One program repeatedly guessed six different integer values for the six vertices in the middle, then computed the other six from the solution of a system of linear equations.
The integer at an outer vertex x is computed as x=(26+a+b-(c+d+e+f))/2, where c,d,e,f are the inner vertices closest to x, and a,b are the inner vertices farthest from x.
7 = (26 + 5+11 - (4+6+8+10)) /2
2 = (26 + 6+5 - (4+8+10+11))/2
It succeeded when the resulting twelve integers were exactly (1,2,...,12) in some order.
Another program used "hill climbing" techniques: place the twelve integers (1,2,...,12) on the vertices randomly, then repeatedly find and execute a "beneficial trade": a pair of integers such that when their positions are traded, some "figure of merit" is improved. In this case the figure of merit was the mean square error of the line sums, namely the sum, over the six lines, of the squares of the differences between their current line sum and the desired line sum of 26.
1. Find all the combinations of 4 of the first 12 integers which sum to 26. Of the 495 possible (12 items taken 4 at a time), 33 sum to 26.
2. For the set of 33, it then generated all the groups of 6 where each integer is used exactly twice. There are 514.
3. It then searched those 514 for the cases where there is only one (or none) shared values been every pair of 6 sums. There are 20. I then randomly picked one of the 20, and in a few minutes (manually) juggled the placement on the star.
Rock wrote (try saying that with a mouth full of gumbo), "
this thred makes my brain hurt..."
Welcome to the club!