cp's OEIS Frontend

The gcd is given in A003989, and v is given in A345416. Minimal means minimize u^2+v^2. We follow Maple, PARI, etc., in setting u=0 and v=1 when m=n. If we ignore the diagonal, the v table is the transpose of the u table.

The old definition was: Triangle T(a,b) read by rows giving number of steps in simple Euclidean algorithm for gcd(a,b) (a > b >= 1). [For this, see A049834.]

For example <11,3> -> <8,3> -> <5,3> -> <3,2> -> <2,1> -> <1,1> -> <1,0> takes 6 steps.

The number of steps function can be defined inductively by T(a,b) = T(b,a), T(a,0) = 0, and T(a+b,b) = T(a,b)+1.

The simple Euclidean algorithm is the Euclidean algorithm without divisions. Given a pair of positive integers with a>=b, let = . This is iterated until a^(m)=0. Then T(a,b) is the number of steps m.

Note that row n starts at k = 1; the number of steps to compute gcd(n,0) or gcd(0,n) is not shown. - T. D. Noe, Oct 29 2007

(Conjecture) Let N=2*n and k=1,...,n. Let A_{N,0}, A_{N,1}, ..., A_{N,n-1} be the n X n unit-primitive matrices (see [Jeffery]) associated with N. Define the Chebyshev polynomials of the second kind by the recurrence U_0(x)=1, U_1(x)=2*x and U_r(x)=2*x*U_(r-1)(x)-U_(r-2)(x) (r>1). Define the column vectors V_(k-1)=(U_(k-1)(cos(Pi/N)), U_(k-1)(cos(3*Pi/N)), ..., U_(k-1)(cos((2*n-1)*Pi/N)))^T, where B^T denotes the transpose of matrix B. Let S_N=[V_0,V_1,...,V_(n-1)] be the n X n matrix formed by taking the components of vector V_(k-1) as the entries in column k-1 (V_(k-1) gives the ordered spectrum of A_{N,k-1}). Let X_N=[S_N]^T*S_N, and let [X_N](i,j) denote the entry in row i and column j of X_N, i,j in {0,...,n-1}. Then also T(n,k)=[X_N](k-1,k-1); that is, row n of the triangle is given by the main diagonal entries of X_N. Hence T(n,k) is the sum of squares T(n,k) = sum[m=1,...,n (U_(k-1)(cos((2*m-1)*Pi/N)))^2]=[V_(k-1)]^T*V_(k-1). - L. Edson Jeffery, Jan 20 2012

Conjecture that antidiagonal sums are A023855. - L. Edson Jeffery, Jan 20 2012

Viewed as a sequence of rows, consider the subsequences (of rows) that contain every positive integer. The lexicographically latest of these subsequences consists of the rows with row numbers in A066680 U {1}; this is the only one that contains its own row numbers only once. - Peter Munn, Dec 04 2019

A198435(n) = A198384(a(n)); A198439(n) = A198388(a(n));

A198436(n) = A198385(a(n)); A198440(n) = A198389(a(n));

A198437(n) = A198386(a(n)); A198441(n) = A198390(a(n));

A198438(n) = A198387(a(n)).

From Robert G. Wilson v, May 10 2002: (Start)

The first term of the k-th row is k. The first leading diagonal contains all 1's. The second leading diagonal contains twice the triangular numbers = n*(n-1).

For p prime, the sum of the p-th row is (p^3 - p^2 + 2)/2.

Proof: The p-th row is p, 2*p, 3*p, ..., (p-2)*p, (p-1)*p, 1. The sum of the row = p*(1 + 2 + 3 + ... + (p-2) + (p-1)) + 1 = p*(p-1)*p/2 + 1 = (p^3 - p^2 + 2)/2. (End) [Edited by Petros Hadjicostas, May 27 2020]

In the square array where T(i,j) = T(j,i), the natural extension of the triangle, each set of rows and columns with common indices [d1, d2, ..., ds] define a group multiplication table on their grid, if the d1, d2, ..., ds are the set of divisors of a squarefree number [A. Jorza]. - R. J. Mathar, May 03 2007

T(n,k) is the minimum number of squares necessary to fill a rectangle with sides of length n and k. - Stefano Spezia, Oct 06 2018

The gcd is 1 unless m=n when it is m; v is given in A345418. Minimal means minimize u^2+v^2. We follow Maple, PARI, etc., in setting u=0 and v=1 when m=n. If we ignore the diagonal, the v table is the transpose of the u table.