Let H be an infinite-dimensional real Hilbert space, let
d>0, and suppose that S is a set of points (not necessarily
countable) in H such that the distance between any two
distinct points in S is equal to d. Show that there is a point
y∈H such that
{d2(x−y):x∈S}
is an orthonormal system of vectors in H.
Solution (official)
It is clear that, if B is an orthonormal system in a Hilbert space
H, then {(d/2)e:e∈B} is a set of
points in H, any two of which are at distance d apart.
We need to show that every set S of equidistant points is a
translate of such a set.
We begin by noting that if x1,x2,x3,x4∈S are four
distinct points, then
⟨x2−x1,x2−x1⟩⟨x2−x1,x3−x1⟩⟨x2−x1,x4−x3⟩=d2,=21(∥x2−x1∥2+∥x3−x1∥2−∥x2−x3∥2)=2d2,=⟨x2−x1,x4−x1⟩−⟨x2−x1,x3−x1⟩=2d2−2d2=0.
This shows that scalar products among vectors which are finite linear
combinations of the form
λ1x1+λ2x2+⋯+λnxn,
where x1,x2,…,xn are distinct points in S and
λ1,λ2,…,λn are integers with
λ1+λ2+⋯+λn=0, are universal
across all such sets S in all Hilbert spaces H; in
particular, we may conveniently evaluate them using examples of our
choosing, such as the canonical example above in Rn. In
fact this property trivially follows also when coefficients
λi are rational, and hence by continuity any real numbers
with sum 0.
If S={x1,x2,…,xn} is a finite set, we form
x=n1(x1+x2+⋯+xn), pick a non-zero vector
z∈[span(x1−x,x2−x,…,xn−x)]⊥
and seek y in the form y=x+λz for a suitable
λ∈R. We find that
⟨x1−y,x2−y⟩=⟨x1−x−λz,x2−x−λz⟩=⟨x1−x,x2−x⟩+λ2∥z∥2.⟨x1−x,x2−x⟩ may be computed by our remark
above as
⟨x1−x,x2−x⟩=2d2⟨(nn−1,−n1,−n1,…,−n1)⊤,(−n1,nn−1,−n1,…,−n1)⊤⟩Rn=2d2(−n22(n−1)+n2n−2)=−2nd2.
So the choice λ=2n∥z∥d will make all
vectors d2(xi−y) orthogonal to each other; it is
easily checked as above that they will also be of length one.
Let now S be an infinite set. Pick an infinite sequence
T={x1,x2,…,xn,…} of distinct points in S. We
claim that the sequence
yn=n1(x1+x2+⋯+xn)
is a Cauchy sequence in H. (This is the crucial
observation.) Indeed, for m>n, the norm ∥ym−yn∥ may be
computed by the above remark as
∥ym−yn∥2=2d2(m1−n1,…,m1−n1,m1,…,m1)⊤Rm2=2d2(m2n2n(m−n)2+m2m−n)=2d2⋅mnm−n=2d2(n1−m1)→0,m,n→∞.
By completeness of H, it follows that there exists a
limit
y=n→∞limyn∈H.
We claim that y satisfies all conditions of the problem. For
m>n>p, with n,p fixed, we compute
∥xn−ym∥2=2d2(−m1,…,−m1,1−m1,−m1,…,−m1)⊤Rm2=2d2(m2m−1+m2(m−1)2)=2d2⋅mm−1→2d2,m→∞,
showing that ∥xn−y∥=d/2, as well as
⟨xn−ym,xp−ym⟩=2d2(m2m−2−m2(1−m1))=−2md2→0,m→∞,
showing that ⟨xn−y,xp−y⟩=0, so that
{d2(xn−y):n∈N}
is indeed an orthonormal system of vectors.
This completes the proof in the case when T=S, which we can
always take if S is countable. If it is not, let x′,x′′ be any
two distinct points in S∖T. Then applying the above
procedure to the set
T′={x′,x′′,x1,x2,…,xn,…} it follows that the
limit
y=n→∞limn+2x′+x′′+x1+x2+⋯+xn=n→∞limnx1+x2+⋯+xn=y
satisfies that
{d2(x′−y),d2(x′′−y)}∪{d2(xn−y):n∈N}
is still an orthonormal system.
This it true
for any distinct x′,x′′∈S∖T; it follows that the
entire system
{d2(x−y):x∈S}
is an orthonormal system of vectors in H, as required.
How the field did
contestants scored
255
average (of 20)
1.39
solved (≥ 80%)
3.9%
near-0 (≤ 10%)
90.2%
discrimination
0.45
Score distribution (field cohort)
Computed on contestants with a meaningful total (field cohort); discrimination is the corrected item–total correlation.