login
Triangle read by rows: T(n,k), for k=-n..n-1, is the scaled (by 2^n n!) probability that the sum of n uniform [-1, 1] variables is between k and k+1.
4

%I #38 Dec 21 2022 04:35:22

%S 1,1,1,3,3,1,1,7,16,16,7,1,1,15,61,115,115,61,15,1,1,31,206,626,1056,

%T 1056,626,206,31,1,1,63,659,2989,7554,11774,11774,7554,2989,659,63,1,

%U 1,127,2052,13308,47349,105099,154624,154624,105099,47349,13308,2052,127,1,1,255,6297,56935,274677

%N Triangle read by rows: T(n,k), for k=-n..n-1, is the scaled (by 2^n n!) probability that the sum of n uniform [-1, 1] variables is between k and k+1.

%C Equivalently, T(n,k)/n! is the n-dimensional volume of the portion of the n-dimensional hypercube [-1,1]^n cut by the (n-1)-dimensional hyperplanes x_1 + x_2 + ... x_n = k, x_1 + x_2 + ... x_n = k+1.

%C The analogous triangle for the interval [0,1] is that of the Eulerian numbers, A008292.

%C This is the distribution of the flag-descent (fdes) statistic in signed permutations, as introduced by Adin, Brenti, Roichman. The link between fdes and portions of hypercubes follows an argument adapted from Stanley. [Matthieu Josuat-Vergès, Apr 25 2011]

%C T(n,n) = A011818(n), the normalized volume of center slice of n-dimensional cube. - _Paul D. Hanna_, Jan 03 2017

%D Peter Doyle, Myths about card shuffling, talk given at DIMACS Workshop on Puzzling Mathematics and Mathematical Puzzles: a Gathering in Honor of Peter Winkler's 60th Birthday, Rutgers University, Jun 08, 2007

%H Paul D. Hanna, <a href="/A101842/b101842.txt">Table of n, a(n) for n = 1..2070 for rows 1..45 of flattened triangle</a>

%H R. M. Adin, F. Brenti and Y. Roichman, <a href="https://doi.org/10.1006/aama.2001.0731">Descent numbers and major indices for the hyperoctahedral group</a>, Adv. Appl. Math. 27 (2001), 210-224.

%H R. Stanley, <a href="http://math.mit.edu/~rstan/pubs/pubfiles/34a.pdf">Eulerian partitions of a unit hypercube</a>, in Higher Combinatorics (M. Aigner, ed.), Reidel, Dordrecht/Boston, 1977, p. 49.

%F T(n,k) = (n-k) * T(n-1,k-1) + T(n-1,k) + (n+k+1)*T(n-1,k+1).

%F With the column index k starting at 0 the table entries appear to be given by t(n,k) = Sum_{i = 0..k} A173018(n,i)*binomial(n,k-i), n >= 1. - _Peter Bala_, Jul 17 2012

%F E.g.f.: -1 + (1-y) / (1 - y*exp( (1-y^2)*x )). - _Paul D. Hanna_, Jan 03 2017

%F p(n, x) = n! * (x - 1)^n * [z^n] (1 - x)/(exp((x + 1)*z) - x) defines a sequence of polynomials. p(n, x) = (x + 1)^n*A(n, x) where A(n, x) are the Eulerian polynomials as defined by A008292. The coefficients of these polynomials are the T(n, k) if enumerated for n >= 0 and 0 <= k <= 2*n - 1 if n > 0 and T(0,0) = 1. - _Peter Luschny_, Jun 24 2019

%e Triangle of T(n,k), k=-n..n-1 begins:

%e 1, 1

%e 1, 3, 3, 1

%e 1, 7, 16, 16, 7, 1

%e 1, 15, 61, 115, 115, 61, 15, 1

%e 1, 31, 206, 626, 1056, 1056, 626, 206, 31, 1

%e 1, 63, 659, 2989, 7554, 11774, 11774, 7554, 2989, 659, 63, 1

%e The sequence of polynomials starts:

%e p(0, x) = 1

%e p(1, x) = x + 1

%e p(2, x) = x^3 + 3*x^2 + 3*x + 1

%e p(3, x) = x^5 + 7*x^4 + 16*x^3 + 16*x^2 + 7*x + 1

%e p(4, x) = x^7 + 15*x^6 + 61*x^5 + 115*x^4 + 115*x^3 + 61*x^2 + 15*x + 1

%p gf := (1 - x)/(exp((x + 1)*z) - x): ser := series(gf, z, 12):

%p for n from 0 to 6 do p := simplify(n!*(x - 1)^n*coeff(ser,z, n));

%p print(PolynomialTools:-CoefficientList(p, x)) od: # _Peter Luschny_, Jun 24 2019

%t T[n_, k_] /; -n <= k <= n-1 := T[n, k] = (n-k) T[n-1, k-1] + T[n-1, k] + (n+k+1) T[n-1, k+1]; T[1, -1] = T[1, 0] = 1; T[_, _] = 0;

%t Table[T[n, k], {n, 1, 8}, {k, -n, n-1}] (* _Jean-François Alcover_, Jun 27 2019 *)

%o (PARI) {T(n,k) = polcoeff( n!*polcoeff( -1 + (1-y) / (1 - y*exp( (1-y^2)*x +x*O(x^n) )) ,n,x),k,y)}

%o for(n=1,10,for(k=1,2*n,print1(T(n,k),", "));print("")) \\ _Paul D. Hanna_, Jan 03 2017

%Y Cf. A101845, A102012, A011818, A173018.

%Y See also A008292 (Eulerian numbers).

%K nonn,tabf

%O 1,4

%A _David Applegate_, based on Peter Doyle's talk, Jun 10 2007

%E Terms a(57) and beyond from _Paul D. Hanna_, Jan 03 2017