login
1/4 of the total number of integral quadruples with sum = n and sum of squares = n^2.
5

%I #45 Mar 09 2023 04:51:59

%S 1,2,4,2,7,8,7,2,13,14,13,8,13,14,28,2,19,26,19,14,28,26,25,8,37,26,

%T 40,14,31,56,31,2,52,38,49,26,37,38,52,14,43,56,43,26,91,50,49,8,49,

%U 74,76,26,55,80,91,14,76,62,61,56,61,62,91,2,91,104,67,38,100,98,73,26,73,74,148,38,91,104,79,14,121,86,85,56

%N 1/4 of the total number of integral quadruples with sum = n and sum of squares = n^2.

%C If instead we count only primitive quadruples (meaning quadruples (h,i,j,k) with gcd(h,i,j,k) = 1) we get A278085(n).

%C Conjectures from _Colin Mallows_, Jun 12 2022: (Start)

%C Given a natural number n, a "quad" for n is a quadruple q = (h,i,j,k) of integers with sum(q) = h+i+j+k = n and sum(q^2) = h^2+i^2+j^2+k^2 = n^2.

%C A quad q is "primitive" if gcd(h,i,j,k) = 1. Define pq(n) = A278085(n) to be the number of distinct primitive quads for n, and tq(n) (the present sequence) to be the total number of quads for n.

%C Conjecture 1: (Based on the data for n <= 5000) pq/4 and tq/4 are multiplicative sequences.

%C Conjecture 2: When n = p^k, p prime and k >= 1:

%C if p = 2, k = 1 then pq(q)/4 = 1 and tq(n)/4 = 2;

%C if p = 2, k >= 2 then pq(q)/4 = 0 and tq(n)/4 = 2;

%C if p = 3, k >= 1 then pq(q)/4 = n and tq(n)/4 = (3*n-1)/2;

%C if p == 5 (mod 6), k >= 1 then pq(q)/4 = (p+1)*n/p and tq(n)/4 = n + 2*(n-1)/(p-1);

%C if p == 1 (mod 6), k >= 1 then pq(q)/4 = (p-1)*n/p and tq(n)/4 = n.

%C (End)

%C Conjecture: the numbers n for which a(n) = n have a positive asymptotic density.

%H Robert Israel, <a href="/A354766/b354766.txt">Table of n, a(n) for n = 1..650</a>

%e Solutions for n = 1: (1,0,0,0) and all permutations thereof.

%e n=2: (2,0,0,0) and (1,1,1,-1).

%e n=3: (3,0,0,0) and (2,2,-1,0).

%e n=4: (4,0,0,0) and (2,2,2,-2). Eight solutions, so a(4) = 8/4 = 2. None are primitive, so A278085(4) = 0.

%e n=5: (5,0,0,0) and (4,2,-2,1). 4+24 solutions, so a(5) = 28/4 = 7. 24 are primitive, so A278085(5) = 24/4 = 6.

%p f:= proc(n) local d; add(g3(n-d, n^2 - d^2), d=-n .. n)/4 end proc:

%p g3:= proc(x,y) option remember; local m,c;

%p if x^2 > 3*y then return 0 fi;

%p m:= floor(sqrt(y));

%p add(g2(x-c,y - c^2), c=- m.. m)

%p end proc:

%p g2:= proc(x,y) option remember;

%p local v;

%p v:= 2*y - x^2;

%p if not issqr(v) then 0

%p elif v = 0 then 1

%p else 2

%p fi

%p end proc:

%p map(f, [$1..100]); # _Robert Israel_, Feb 16 2023

%t f[n_] := Sum[g3[n - d, n^2 - d^2], {d, -n, n}]/4 ;

%t g3[x_, y_] := g3[x, y] = Module[{m}, If[x^2 > 3*y, 0, m = Floor[Sqrt[y]]; Sum[g2[x - c, y - c^2], {c, -m, m}]]];

%t g2[x_, y_] := g2[x, y] = Module[{v}, v = 2*y - x^2; Which[!IntegerQ@Sqrt[v], 0, v == 0, 1, True, 2]];

%t f /@ Range[100] (* _Jean-François Alcover_, Mar 09 2023, after _Robert Israel_ *)

%Y Cf. A278085, A354777, A354778.

%Y See also A353589 (counts nondecreasing nonnegative (h,i,j,k) such that (+-h, +-i, +-j, +-k) is a solution).

%K nonn,look

%O 1,2

%A _N. J. A. Sloane_, Jun 19 2022, based on an email from _Colin Mallows_, Jun 12 2022