278 CHAPTER 11. INNER PRODUCT SPACES

then for a given function f ∈ X, the function from M which is closest to f in this innerproduct norm is

g =

n∑k=−n

(f, ek) ek

In this case (f, ek) =1√2π

∫ π

−πf (x) eikxdx. These are the Fourier coefficients. The above is

the nth partial sum of the Fourier series.

To show how this kind of thing approximates a given function, let f (x) = x2. Let

M = span

({1√2πe−ikx

}3

k=−3

). Then, doing the computations, you find the closest point

is of the form

1

3

√2π

52

(1√2π

)+

3∑k=1

((−1)

k2

k2

)√2√π

1√2πe−ikx +

3∑k=1

((−1)

k2

k2

)√2√π

1√2πeikx

and now simplify to get

1

3π2 +

3∑k=1

(−1)k

(4

k2

)cos kx

Then a graph of this along with the graph of y = x2 is given below. In this graph, the dashedgraph is of y = x2 and the solid line is the graph of the above Fourier series approximation.

If we had taken the partial sum up to n much bigger, it would have beenvery hard to distinguish between the graph of the partial sum of theFourier series and the graph of the function it is approximating. Thisis in contrast to approximation by Taylor series in which you only getapproximation at a point of a function and its derivatives. These arevery close near the point of interest but typically fail to approximatethe function on the entire interval.

11.3 Riesz Representation Theorem

The next theorem is one of the most important results in the theory of inner product spaces.It is called the Riesz representation theorem.

Theorem 11.3.1 Let f ∈ L (X,F) where X is an inner product space of dimension n.Then there exists a unique z ∈ X such that for all x ∈ X,

f (x) = (x, z) .

Proof: First I will verify uniqueness. Suppose zj works for j = 1, 2. Then for all x ∈ X,

0 = f (x)− f (x) = (x, z1 − z2)

and so z1 = z2.It remains to verify existence. By Lemma 11.2.1, there exists an orthonormal basis,

{uj}nj=1 . If there is such a z, then you would need f (uj) = (uj , z) and so you would need

f (uj) = (z, uj) . Also you must have z =∑

i (z, uj)uj . Therefore, define

z ≡n∑

j=1

f (uj)uj .

278 CHAPTER 11. INNER PRODUCT SPACESthen for a given function f € X, the function from M which is closest to f in this innerproduct norm isnI= S- (f, en) ekk=—-nIn this case (f,ex) = Te i f (x) e**dx. These are the Fourier coefficients. The above isthe n*” partial sum of the hone series.To show how this kind of thing approximates a given function, let f (2) = 27. Let43M = span {ae} ) . Then, doing the computations, you find the closest pointk=-3Jonis of the formje (Ze) (Q 1)" *) Vai etn (SEk=1 k=11 .J2 7 eike2) V2rand now simplify to get31gt "(a 5) 605kThen a graph of this along with the graph of y = 2? is given below. In this graph, the dashedgraph is of y = x? and the solid line is the graph of the above Fourier series approximation.If we had taken the partial sum up to n much bigger, it would have been/ very hard to distinguish between the graph of the partial sum of theFourier series and the graph of the function it is approximating. Thisis in contrast to approximation by Taylor series in which you only get0 7 tk approximation at a point of a function and its derivatives. These are4 very close near the point of interest but typically fail to approximatethe function on the entire interval.Ko211.3. Riesz Representation TheoremThe next theorem is one of the most important results in the theory of inner product spaces.It is called the Riesz representation theorem.Theorem 11.3.1 Let f € £(X,F) where X is an inner product space of dimension n.Then there exists a unique z € X such that for alla € X,f (#) = (@,2).Proof: First I will verify uniqueness. Suppose z; works for 7 = 1,2. Then for all x € X,O= f(a) — f(a) = (a, 21 — 22)and so z1 = 22.It remains to verify existence. By Lemma 11.2.1, there exists an orthonormal basis,{u; Yea . If there is such a z, then you would need f (u;) = (u,;,z) and so you would needf (uj) = (¢,u;). Also you must have z = 57; (z,u;) uj. Therefore, definej=1