A359364 Triangle read by rows. The Motzkin triangle, the coefficients of the Motzkin polynomials. M(n, k) = binomial(n, k) * CatalanNumber(k/2) if k is even, otherwise 0.
1, 1, 0, 1, 0, 1, 1, 0, 3, 0, 1, 0, 6, 0, 2, 1, 0, 10, 0, 10, 0, 1, 0, 15, 0, 30, 0, 5, 1, 0, 21, 0, 70, 0, 35, 0, 1, 0, 28, 0, 140, 0, 140, 0, 14, 1, 0, 36, 0, 252, 0, 420, 0, 126, 0, 1, 0, 45, 0, 420, 0, 1050, 0, 630, 0, 42, 1, 0, 55, 0, 660, 0, 2310, 0, 2310, 0, 462, 0
Offset: 0
Examples
Triangle M(n, k) starts: [0] 1; [1] 1, 0; [2] 1, 0, 1; [3] 1, 0, 3, 0; [4] 1, 0, 6, 0, 2; [5] 1, 0, 10, 0, 10, 0; [6] 1, 0, 15, 0, 30, 0, 5; [7] 1, 0, 21, 0, 70, 0, 35, 0; [8] 1, 0, 28, 0, 140, 0, 140, 0, 14; [9] 1, 0, 36, 0, 252, 0, 420, 0, 126, 0;
Links
- M. Artioli, G. Dattoli, S. Licciardi, and S. Pagnutti, Motzkin Numbers: an Operational Point of View, arXiv:1703.07262 [math.CO], 2017, (Table 1 on p. 3).
Crossrefs
Programs
-
Maple
CatalanNumber := n -> binomial(2*n, n)/(n + 1): M := (n, k) -> ifelse(irem(k, 2) = 1, 0, CatalanNumber(k/2)*binomial(n, k)): for n from 0 to 9 do seq(M(n, k), k = 0..n) od; # Alternative, as coefficients of polynomials: p := n -> hypergeom([(1 - n)/2, -n/2], [2], (2*x)^2): seq(print(seq(coeff(simplify(p(n)), x, k), k = 0..n)), n = 0..9); # Using the exponential generating function: egf := exp(x)*BesselI(1, 2*x*t)/(x*t): ser:= series(egf, x, 11): seq(print(seq(coeff(simplify(n!*coeff(ser, x, n)), t, k), k = 0..n)), n = 0..9);
-
Python
from functools import cache @cache def M(n: int, k: int) -> int: if k % 2: return 0 if n < 3: return 1 if n == k: return (2 * (n - 1) * M(n - 2, n - 2)) // (n // 2 + 1) return (M(n - 1, k) * n) // (n - k) for n in range(10): print([M(n, k) for k in range(n + 1)])
Formula
Let p(n, x) = hypergeom([(1 - n)/2, -n/2], [2], (2*x)^2).
M(n, k) = [x^k] p(n, x).
M(n, k) = [t^k] (n! * [x^n] exp(x) * BesselI(1, 2*t*x) / (t*x)).
M(n, k) = [t^k][x^n] ((1 - x - sqrt((x-1)^2 - (2*t*x)^2)) / (2*(t*x)^2)).
M(n, n) = A126120(n).
M(n, n-1) = A138364(n), the number of Motzkin n-paths with exactly one flat step.
M(2*n, 2*n) = A000108(n), the number of peakless Motzkin paths having a total of n up and level steps.
M(4*n, 2*n) = A359647(n), the central terms without zeros.
M(2*n+2, 2*n) = A002457(n) = (-4)^n * binomial(-3/2, n).
Sum_{k=0..n} M(n - k, k) = A023426(n).
Sum_{k=0..n} k * M(n, k) = 2*A014531(n-1) = 2*GegenbauerC(n - 2, -n, -1/2).
Sum_{k=0..n} i^k*M(n, k) = A343773(n), (i the imaginary unit), is the excess of the number of even Motzkin n-paths (A107587) over the odd ones (A343386).
Sum_{k=0..n} Sum_{j=0..k} M(n, j) = A189912(n).
Sum_{k=0..n} Sum_{j=0..k} M(n, n-j) = modified A025179(n).
For a recursion see the Python program.
Comments