This is a front-end for the Online Encyclopedia of Integer Sequences, made by Christian Perfect. The idea is to provide OEIS entries in non-ancient HTML, and then to think about how they're presented visually. The source code is on GitHub.
%I A283184 #48 Mar 02 2024 12:28:48 %S A283184 1,0,2,14,122,1262,15466,219646,3551194,64431374,1296712778, %T A283184 28672204574,691007296954,18029138380846,506320912190506, %U A283184 15228632768870462,488405396197019546,16638380026019579726,600022595692147574794,22836184629309211495774,914717041435012519583098 %N A283184 a(n) is the number of symmetric permutations (p(1),p(2),...,p(m)) of (1,2,...,m), m=2n or m=2n+1, with p(m+1-k) = m+1-p(k) for 1<=k<=m, such that adjacent numbers do not differ by 1. a(n) is also the number of point-symmetric arrangements of m non-attacking kings on an m X m board, with one in each row and column. %C A283184 For m=2n+1 the symmetry requires p(n+1)=n+1. That is why the number of permutations is the same for m=2n and m=2n+1. %C A283184 The n-th element of any permutation is not allowed to be n because otherwise the next element would be n+1. Because of the symmetry it is sufficient to consider the first n elements. Any such n-tuple can be created by a permutation of length n, last element smaller than n: Each element b(k) > n has to be replaced by m+1-b(k). %C A283184 Example m=6: Original symmetric permutation 536142, 3-tuple 536 created by 231: 5 is replaced by 7-5 and 6 by 7-6. %C A283184 How many such n-tuples can be created by a n-permutation? %C A283184 Let us analyze the example above: There are two pairs of adjacent numbers (23 and 31) in the permutation 231. The difference of the first pair is 1, so either 2 or 3 must be replaced, whereas the second pair represents a "gap" (difference > 1), so that 1 can be kept or replaced by 6. %C A283184 This way, 231 creates four 3-tuples: 531, 241, 536, 246. %C A283184 Let generally g be the number of gaps in a n-permutation (0<=g<=n-1). Then the number of related n-tuples is 2^(g+1) because the first element of the permutation and each element behind a gap can be arbitrarily replaced or not. On the other hand, when the first element of a section between successive gaps is selected, there is no choice for the replacement of the other elements. %C A283184 When q is a n-permutation, the number of gaps is g(q) = Sum_{j=1..n-1} sign(|p(j+1)-p(j)|-1). (sign = signum) %C A283184 The extension up to n=50 was done by a new algorithm, see link "Fast recurrence". - _Gerhard Kirchner_, Mar 17 2017 %H A283184 Alois P. Heinz, <a href="/A283184/b283184.txt">Table of n, a(n) for n = 0..400</a> (terms n = 0..50 from Gerhard Kirchner) %H A283184 Gerhard Kirchner, <a href="/A283184/a283184.txt">Fast recurrence</a> %F A283184 Let q be any permutation (p(1), p(2),..., p(n)) with p(n) < n and g(q) = Sum_{j=1..n-1} sgn(|p(j+1)-p(j)|-1). %F A283184 a(n) = Sum_{each q} 2^(g(q)+1). %F A283184 a(n) ~ exp(-1) * 2^n * n!. - _Vaclav Kotesovec_, Apr 20 2017 %e A283184 Example 1, m=5: %e A283184 The matrix, transforming 12345 into 41352, can also be thought of as a chessboard; each "1" is a king. %e A283184 ./0 0 0 1 0\ /1\ /4\ %e A283184 | 1 0 0 0 0 | |2| |1| %e A283184 | 0 0 1 0 0 |*|3|=|3| %e A283184 | 0 0 0 0 1 | |4| |5| %e A283184 .\0 1 0 0 0/ \5/ \2/ %e A283184 Example 2, m=6: %e A283184 q is a 3-permutation not ending on 3: %e A283184 q g(q) 2^(g(q)+1) Symmetric 6-permutations, |p(j+1)-p(j)|>1 %e A283184 132 1 4 135246, 635241, 142536, 642531 %e A283184 231 1 4 531642, 536142, 241635, 246135 %e A283184 312 1 4 315264, 362514, 462513, 415263 %e A283184 321 0 2 426153, 351624 %e A283184 Result: a(3)=14. %p A283184 b:= proc(n, s, l) option remember; `if`(s={}, %p A283184 `if`(abs(n/2-l)<1, 0, 1), add(add(`if`(abs(j-l)=1, 0, %p A283184 b(n, s minus {i}, i)), j=[i, n-i]), i=s)) %p A283184 end: %p A283184 a:= n-> b(2*n+1, {$1..n}, -1): %p A283184 seq(a(n), n=0..10); # _Alois P. Heinz_, Mar 15 2017 %p A283184 # second Maple program: %p A283184 a:= proc(n) option remember; `if`(n<4, (n-1)* %p A283184 (7*n^2-5*n-6)/6, (2*n+1)*a(n-1) -(2*n-5)* %p A283184 (a(n-2)+a(n-3)) +(2*n-6)*a(n-4)) %p A283184 end: %p A283184 seq(a(n), n=0..20); # _Alois P. Heinz_, Mar 17 2017 %t A283184 a[n_] := a[n] = If[n<4, (n-1)*(7n^2-5n-6)/6, (2n+1)*a[n-1] - (2n-5)*(a[n-2] + a[n-3]) + (2n-6)*a[n-4]]; Table[a[n], {n, 0, 20}] (* _Jean-François Alcover_, Mar 18 2017, after _Alois P. Heinz_ *) %o A283184 (Visual Basic) %o A283184 a(n) = fusum(n, 1, "", "", 0, 0) with %o A283184 Function fusum(n, t, permu$, pile$, g, su) %o A283184 If t = n + 1 Then %o A283184 su = su + 2 ^ (g + 1) %o A283184 Else %o A283184 la = n + 1 - t %o A283184 If t = 1 Then %o A283184 la = n - 1 %o A283184 For k = 1 To n: pile$ = pile$ + Chr(k): Next %o A283184 End If %o A283184 For s = la To 1 Step -1 %o A283184 y = Asc(Mid(pile$, s)) %o A283184 If t = 1 Then ad = 0 Else ad = Sgn(Abs(y - Asc(permu$)) - 1) %o A283184 fusum = fusum(n, t + 1, Chr(y) + permu$, Left(pile$, s - 1) + Mid(pile$, s + 1), g + ad, su) %o A283184 Next %o A283184 If t = n Then fusum = su %o A283184 End If %o A283184 End Function %o A283184 (PARI) a(n)={subst(serlaplace(polcoef((1 - x)/(1 + (1 - 2*y)*x + 2*y*x^2) + O(x*x^n), n)), y, 1)} \\ _Andrew Howroyd_, Mar 01 2024 %Y A283184 Cf. A002464. %K A283184 nonn %O A283184 0,3 %A A283184 _Gerhard Kirchner_, Mar 02 2017 %E A283184 a(14)-a(20) from _Alois P. Heinz_, Mar 15 2017