triangles
with prime side lengths and vertices on lattice
points.
On April 21, Oscar Volpatti sent the following
puzzle.
For this puzzle, the only valid triangles are
those whose three side lengths are prime numbers
p,q,r (not necessarily distinct).
The
task is to find valid triangles embedded in a
given lattice; in other words, triangles whose
three vertices are lattice points.
An
infinite family of lattices is available: for
each positive integer n there is a
"n-dimensional integer lattice", denoted as Z^n,
which is the set of all points with integer
coordinates in n-dimensional Euclidean space.
Euclidean distance between lattice points
A=(a_1,...a_n) and B=(b_1,...,b_n) is computed
as usual: d(A,B) = sqrt((a_1-b_1)^2 + ... +
(a_n-b_n)^2).
The vertices of a proper
triangle do not all lie on the same straight
line, so lattice Z^1 must be discarded and the
first few lattices to be considered are the
"square lattice" Z^2 on the plane and the "cubic
lattice" Z^3 in 3D space.
Answer the
following questions.
Q1) Which
valid triangles are embeddable in Z^2? Q2)
Which ones are embeddable in Z^3 but not in Z^2?
Q3) Which ones are embeddable in Z^4 but not in
Z^3?
Q4) Which
valid equilateral triangles are embeddable in
some lattice Z^n? Q5) Which valid triangles
are embeddable in no lattice Z^n, no matter
which value of n is chosen? Q6) Is there some
upper bound nmax such that any valid triangle
embeddable in some lattice Z^n is also
guaranteed to be embeddable in Z^nmax?
Some (non-valid) examples.
The rectangular triangle with side lengths 3,4,5
is easily embeddable in Z^2, by choosing
vertices: {(0,0),(4,0),(0,3)}. The
isosceles triangle with side lengths 4,7,7 is
not embeddable in Z^2 but it's embeddable in
Z^3, by choosing vertices:
{(0,0,0),(4,0,0),(2,3,6)}.
I hope that you'll find this puzzle interesting.
It was partially inspired by
puzzle 1155, also related to triangle
searching and prime distances.