The Riesz Representation Theorem

MA 466
Kurt Bryan
Let H be a Hilbert space over lR or C
l , and T a bounded linear functional
on H (a bounded operator from H to the field, lR or C l , over which H is
defined). The following is called the Riesz Representation Theorem:
Theorem 1 If T is a bounded linear functional on a Hilbert space H then
there exists some g ∈ H such that for every f ∈ H we have
T (f ) =< f, g > .
Moreover, ∥T ∥ = ∥g∥ (here ∥T ∥ denotes the operator norm of T , while ∥g∥
is the Hilbert space norm of g.
Proof: Let’s assume that H is separable for now. It’s not really much harder
to prove for any Hilbert space, but the separable case makes nice use of the
ideas we developed regarding Fourier analysis. Also, let’s just work over lR.
Since H is separable we can choose an orthonormal basis ϕj , j ≥ 1, for
∑ and set aj = T (ϕj ). Choose f ∈ H,
H. Let T be a bounded linear functional
let cj =< f, ϕj >, and define fn = nj=1 cj ϕj . Since the ϕj form a basis we
know that ∥f − fn ∥ → 0 as n → ∞.
Since T is linear we have
∑
n
T (fn ) = aj cj . (1)
j=1

Since T is bounded, say with norm ∥T ∥ < ∞, we have

|T (f ) − T (fn )| ≤ ∥T ∥∥f − fn ∥. (2)
Because ∥f − fn ∥ → 0 as n → ∞ we conclude from equations (1) and (2)
that
∑
∞
T (f ) = lim T (fn ) = a j cj . (3)
n→∞
j=1

In fact, the sequence aj must itself be square-summable. To see this, first

note that since |T (f )| ≤ ∥T ∥∥f ∥ we have
( ∞ )1/2
∑∞ ∑
cj aj ≤ ∥T ∥ c2j . (4)
j=1 j=1

1
Equation (4) must hold for any square-summable sequence cj (since any such
cj corresponds to some element in H). Fix a positive integer N and define
a sequence cj = aj for j ≤ N , cj = 0 for j > N . Clearly such a sequence is
square-summable and equation (4) then yields
( )1/2
∑
N ∑
N
a2j ≤ ∥T ∥ a2j
j=1 j=1

or ( )1/2
∑
N
a2j ≤ ∥T ∥. (5)
j=1

Thus aj is square-summable, since the sequence of partial sums is bounded

above. ∑
Since aj is square-summable ∑
the function g = j aj ϕj is well-defined as
an element of H, and T (f ) = j aj cj =< f, g >. Finally, equation (5)
makes it clear that ∥g∥ ≤ ∥T ∥. But from Cauchy-Schwarz we also have
|T (f )| = | < f, g > | ≤ ∥f ∥∥g∥ or |T∥f(f∥)| ≤ ∥g∥, implying ∥T ∥ ≤ ∥g∥, so
∥T ∥ = ∥g∥.

An Application

This examples illustrates how functional analytic methods are used in

PDE (though the example is for an ODE). Consider the ODE

−f ′′ (x) + b(x)f (x) = q(x) (6)

on the interval 0 < x < 1, where b(x) ≥ δ > 0 for some δ; assume the
functions b and q are continuous on [0, 1]. We want a solution to equation
(6) with f ′ (0) = f ′ (1) = 0 (other boundary conditions are possible). If we
multiply (6) by a C 1 function ϕ and integrate the first term, −f ′′ ϕ by parts
from x = 0 to x = 1, we obtain
∫ 1 ∫ 1
′ ′
(f (x)ϕ (x) + b(x)f (x)ϕ(x)) dx = q(x)ϕ(x) dx. (7)
0 0

Equation (7) must hold for any ϕ ∈ C 1 ([0, 1]), if f is a C 2 (0, 1) solution
to equation (6) which is continuous on [0, 1]. Conversely, if for a given C 2

2
function f we find that (7) holds for all ϕ, then f must be a solution to
equation (6), for if we “undo” the integration by parts in (7) we obtain

ϕ(1)f ′ (1) − ϕ(0)f ′ (0) + ϕ(x)(−f ′′ (x) + b(x)f (x)) = ϕ(x)q(x)

for all ϕ. A familiar PDE argument then shows that f ′ (0) = f ′ (1) = 0 and
equation (6) must hold.
We’re going to show that there is a unique solution to equation (7). Such a
“solution” will not necessarily be twice-differentiable as required by equation
(6), but it will satisfy equation (7). Equation (7) is often called the “weak”
formulation of the problem.
Define an inner product
∫ 1
< g, h >= (g ′ (x)h′ (x) + b(x)g(x)h(x)) dx
0

on the space C 1 ([0, 1]), and let H denote the completion of the this space.
This is essentially the procedure used on the third problem of the first exam;
the presence of b(x) makes no difference. (Note we must use b ≥ δ > 0 to
√
assure that <, > really is an inner product, so that ∥g∥ = < g, g > = 0 iff
g ≡ 0). The space H is a Hilbert space, and can be interpreted (if need be)
as a subspace of C([0, 1]).
Define a functional T : H → lR by
∫ 1
T (ϕ) = q(x)ϕ(x) dx
0

You can easily check that T is bounded on H (use Cauchy-Schwarz). From

the Riesz Representation Theorem it then follows that there must exist some
function f ∈ H such that
T (ϕ) =< f, ϕ >
for all ϕ ∈ H. This is exactly equation (7), the weak form of the ODE!
The function f that satisfies equation (7) lies in H. Given the conditions
on b (in particular, b ≥ δ > 0 and ∥b∥∞ < ∞ since b ∈ C([0, 1])) the function
f lies in the same space defined in the third problem of the first exam. In
particular, f is a continuous function. To show that f is actually twice-
differentiable requires more work, along with assumptions on the function
q.