Re Permanents of (+1,-1) matrices
Jaap Spies
j.spies at hccnet.nl
Tue Oct 28 18:57:14 CET 2003
N. J. A. Sloane wrote:
> Thanks to everyone who replied. To summarize, let A
> be a nonsingular +1,-1 n X n matrix. The conjecture appears
> to be that for n >= 5 the max permanent of A is given by
> the following (new) sequence:
>
> %I A087981
> %S A087981 0,2,4,24,128,880,6816,60032,589312
> %N A087981 Permanent of the n X n (+1,-1)-matrix with exactly n-1 -1s on the diagonal.
> %C A087981 It is conjectured by Kraeuter and Seifter that for n >= 5 this is the maximum possible value for the permanent of a nonsingular n X n (+1,-1)-matrix with exactly n-1 -1's on the diagonal.
> %C A087981 I don't know for what values of n this has been confirmed. As Edwin Clark points out, it is certainly false for n = 4. - njas
> %C A087981 The maximum possible value for the permanent of a singular n X n (+1,-1)-matrix is obviously n!.
> %D A087891 A. R. Kraeuter and N. Seifter, Some properties of the permanent of (1,-1)-matrices, Linear and Multilinear Algebra 15 (1984), 207-223.
> %D A087981 N. Seifter, Upper bounds for permanents of (1,-1)-matrices, Israel J. Math. 48 (1984), 69-78.
> %D A087891 Edward Tzu-Hsia Wang, On permanents of (1,-1)-matrices, Israel J. Math. 18 (1974), 353-361.
> %O A087981 2,2
> %K A087981 nonn,easy,more
> %A A087981 Gordon Royle (gordon(AT)csse.uwa.edu.au), Oct 28 2003
>
> But what is the max permanent for n <= 4 ? For
> n = 1 2 3 4 we have
> 1 0 6 x where thanks to Ed Clark we know x >= 8
>
> It remains to see if 8 is the max permanent of a 4x4 nonsingular
> +1,-1 matrix - could someone please do that computation?
>
>
> NJAS
>
> .
>
I do have some more. At the moment up to n=25, with my 'new alg' (to be published)
for the permanent:
n = 2: 0
n = 3: 2
n = 4: 4
n = 5: 24
n = 6: 128
n = 7: 880
n = 8: 6816
n = 9: 60032
n = 10: 589312
n = 11: 6384384
n = 12: 75630080
n = 13: 972387328
n = 14: 13483769856
n = 15: 200571078656
n = 16: 3185540657152
n = 17: 53800242216940
n = 18: 962741176500224
n = 19: 18195808235851328
n = 20: 362183230599829504
n = 21: 7572922094356201472
n = 22: 165945771114706763776
n = 23: 3802923921409419771904
n = 24: 90965940198950249168896
n = 25: 2267151124762193502928896
Jaap Spies
More information about the SeqFan
mailing list