mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Puzzles (https://www.mersenneforum.org/forumdisplay.php?f=18)
-   -   Square of Primes (https://www.mersenneforum.org/showthread.php?t=10251)

 davar55 2008-04-30 18:12

Square of Primes

Construct a 5 x 5 square containing distinct primes
such that each row, column and diagonal sums to a distinct prime.

 petrw1 2008-04-30 22:35

[QUOTE=davar55;132471]Construct a 5 x 5 square containing distinct primes
such that each row, column and diagonal sums to a distinct prime.[/QUOTE]

Is it a magic square where every sum is the same?

 davar55 2008-04-30 22:51

A magic square of primes (where every sum is the same) is
solved elsewhere (although it would be a perfectly good
puzzle to re-solve).

Here every sum is a different prime.

 davar55 2008-05-13 18:07

The original problem was perhaps too computationally simple
iota of complexity:

The 25 distinct primes in the square should be the first 25
odd primes {3,5,7,...,97,101}.

(I have a solution which wasn't hard to find by trial and error,
so there must be many solutions; but plain brute force on the 25!
such possible squares is obviously too computationally costly.)

 lavalamp 2008-05-17 02:23

Here's one:[CODE][SPOILER] 239

/
/
/

3, 5, 7, 11, 17 --- 43

13, 19, 29, 23, 43 --- 127

31, 67, 61, 47, 71 --- 277

53, 59, 41, 73, 37 --- 263

79, 83, 89, 97, 101 --- 443

| | | | | \
| | | | | \
| | | | | \

179 233 227 251 269 257[/SPOILER][/CODE]

 S485122 2008-05-17 06:12

There is an error in your calculations : the last row total is of by 6. But the right number is prime so the solution stands :-)

Jacob

 lavalamp 2008-05-17 11:59

Hm, I think I worked the total out right, but wrote it down wrong.

It wasn't just a fluke, honest! ;)

 m_f_h 2008-05-19 14:01

Does the sequence

a(n) = number of square matrices containing the first (2n+1)x(2n+1) odd primes, such that row, column and diagonal sums are distinct primes

exist on OEIS ?

 lavalamp 2008-05-19 14:57

It would appear that there are an awful lot of these out there, so perhaps the challange should be to find a square with the lowest standard deviation of column/row/diagonal totals.

I'll start the ball rolling with a slightly modified version of the last square I posted, with an s.d. of 84.51:[code][SPOILER] 239

/
/
/

3, 5, 7, 11, 17 --- 43

13, 19, 29, 23, 43 --- 127

31, 67, 61, 47, 71 --- 277

53, 59, 89, 73, 37 --- 311

79, 83, 41, 97, 101 --- 401

| | | | | \
| | | | | \
| | | | | \

179 233 227 251 269 257[/SPOILER][/code]

 davar55 2008-05-21 12:54

Here's another solution:

041 005 007 071 003
013 023 029 031 067
059 053 043 047 037
019 011 089 061 017
079 101 083 097 073

Rows: 127,163,239,197,433
Columns: 211,193,251,307,197
Diagonals: 241,167

(Standard deviation: 76.7)

An alternative measure is simply mini-max: minimize the largest sum.
By that measure, lavalamp's solution is a better one.

 All times are UTC. The time now is 16:37.