The M. Riesz extension theorem is a theorem in mathematics, proved by Marcel Riesz during his study of the problem of moments.
Let
E
F\subsetE
K\subsetE
\phi:F\toR
K
K
\phi(x)\geq0 for x\inF\capK.
A linear functional
\psi:E\toR
K
\phi
\phi
\phi
K
\psi|F=\phi and \psi(x)\geq0 for x\inK.
In general, a
K
F
K
E
E=R2, K=\{(x,y):y>0\}\cup\{(x,0):x>0\},
F
x
\phi(x,0)=x
E
However, the extension exists under the additional assumption that
E\subsetK+F,
y\inE,
x\inF
y-x\inK.
The proof is similar to the proof of the Hahn–Banach theorem (see also below).
By transfinite induction or Zorn's lemma it is sufficient to consider the case dim
E/F=1
Choose any
y\inE\setminusF
a=\sup\{\phi(x)\midx\inF, y-x\inK\}, b=inf\{\phi(x)\midx\inF,x-y\inK\}.
We will prove below that
-infty<a\leb
c
a\lec\leb
\psi(y)=c
\psi|F=\phi
\psi
E
\psi
K
z\inK
z=0
z=p(x+y)
z=p(x-y)
p>0
x\inF
z=0
\psi(z)>0
x+y=y-(-x)\inK
\psi(y)=c\geqa\geq\phi(-x)=\psi(-x)
by definition. Thus
\psi(z)=p\psi(x+y)=p(\psi(x)+\psi(y))\geq0.
In the second case,
x-y\inK
\psi(y)=c\leqb\leq\phi(x)=\psi(x)
by definition and so
\psi(z)=p\psi(x-y)=p(\psi(x)-\psi(y))\geq0.
In all cases,
\psi(z)>0
\psi
K
We now prove that
-infty<a\leb
x\inF
y-x\inK
-infty<a
x\inF
x-y\inK
b=infty
b<infty
x\inF
x-y\inK
x\inF
y-x\inK
x'\inF
x'-y\inK
\phi(x)\le\phi(x')
x'-x=(x'-y)+(y-x)\inK
since
K
0\leq\phi(x'-x)=\phi(x')-\phi(x)
since
\phi
K
Let E be a real linear space, and let K ⊂ E be a convex cone. Let x ∈ E/(-K) be such that R x + K = E. Then there exists a K-positive linear functional φ: E → R such that φ(x) > 0.
The Hahn–Banach theorem can be deduced from the M. Riesz extension theorem.
Let V be a linear space, and let N be a sublinear function on V. Let φ be a functional on a subspace U ⊂ V that is dominated by N:
\phi(x)\leqN(x), x\inU.
The Hahn - Banach theorem asserts that φ can be extended to a linear functional on V that is dominated by N.
To derive this from the M. Riesz extension theorem, define a convex cone K ⊂ R×V by
K=\left\{(a,x)\midN(x)\leqa\right\}.
Define a functional φ1 on R×U by
\phi1(a,x)=a-\phi(x).
One can see that φ1 is K-positive, and that K + (R × U) = R × V. Therefore φ1 can be extended to a K-positive functional ψ1 on R×V. Then
\psi(x)=-\psi1(0,x)
is the desired extension of φ. Indeed, if ψ(x) > N(x), we have: (N(x), x) ∈ K, whereas
\psi1(N(x),x)=N(x)-\psi(x)<0,
leading to a contradiction.