I have Ac(0.015) by diagonal sequantal placing nambers 1..N*N from max and min alternating
25 2 22 6 2 17
1  23 5 18  10
24 4  19 9  14
3  20 8  15 12
21 7  16 11 13
but clear mathamatic proof i don't now and this is most interesting. I will try find out such proof and if people will have interest will send this proof to forum. Without proof algotithm hasn't hard foundation.

-

So, what is your algo, svr?
After building your matrix, what are you doing?

Thanks.

Edited by author 17.10.2006 13:53

If we have optimal matrix to take answer is very simple problem. We use double loop, calculate sums A[i,j]+A[i-1,j] and A[i,j-1]+A[i,j] and renew record value searching max.

First number is always : (n - 1).
Second if n even => 3 * ((n - 2) / 2) + 1
If n odd => 3 * ((n - 1) / 2)

Yes, I have also maden by this way. My program is in 7 lines(without var,begin,end - 3).
But I don'y understand why it works in 0.015, except 0.001.
I used only 1 read and if.
strange isn't it?

what is answer if n=8
when only one match with win&lose ,and other is draw
answer should be 6 9 (6(1)+3)?

Edited by author 24.02.2008 13:21

Edited by author 24.02.2008 13:21