2024 SPRING 27200 HOMEWORK 1, DUE 04/01 23:59
1. Basic Definitions, 1 Pt Total
1. Let X be the vector space of all real-valued functions on R, and let E ⊂ X be the subspace of all even
functions. Construct an invertible linear operator from the quotient space X/E to the space of odd functions
on R.
2. Let f be a linear functional on Rn . Prove that there must be some c ∈ Rn such that
n
X
f (x) = ck x k .
k=1
Give a geometric interpretation of the null space Nul(f ).
3. Sketch the picture of the unit sphere for each of the following norms on R2 :
(1) The ℓ1 norm ∥(x, y)∥1 = |x| + |y|.
(2) The ℓ∞ norm ∥(x, y)∥∞ = max |x|, |y| .
(3) A weighted norm
∥(x, y)∥ = max |x|, 2|y| .
4. Consider the convex set
(x − 1)2 + y 2 < 4
in the plane. Sketch the picture of it. Compute the Minkowski functional of this convex set with respect to
0, in terms of the polar coordinate (r, θ).
5. Prove that there does not exist a translation invariant Borel measure on the infinite dimensional space
L2 (R). Here a “Borel measure” on a metric space refers to a measure defined on the σ-algebra generated by
open sets. Thus there does not exist “infinite dimensional Lebesgue measure”. Hint. In this case, the unit
ball can cover infinitely many identical small balls.
2. Exercises from Textbook, 2 Pt Total
1. Chapter 1: Exercise 3 on page 5, Exercise 5 on page 6.
2. Chapter 2: Exercise 2 on page 9.
3. Chapter 3: Exercise 1 and 2 on page 22.
3. Some Geometry Related to Hahn-Banach Theorem, 2 Pt Total
1. As we have mentioned in class, it is not satisfactory if the Hahn-Banach theorem we use in practice relies
on the axiom of choice. Try to prove the Hahn-Banach theorem for Rn without using the axiom of choice.
Hint. In this case, the induction step should terminate after finitely many repetitions.
1
2 2024 SPRING 27200 HOMEWORK 1, DUE 04/01 23:59
2. Let X be a real vector space. Suppose there are linear functionals f and f1 , f2 , · · · , fn , such that
n
\
Nul(fk ) ⊂ Nul(f ).
k=1
Prove that f is in fact a linear combination of f1 , f2 , · · · , fn .
3. Let X be a real vector space. Fix linear functionals f1 , f2 , · · · , fn and a tuple α ∈ Rn . Prove that the
following two properties are equivalent:
• There is some x ∈ X such that
α = f1 (x), · · · , fn (x) .
• If β ∈ Rn is such that
n
X
βk fk = 0,
k=1
then necessarily α · β = 0.
4. Let Q be the subset {x ∈ Rn : xk ≥ 0, ∀k} in Rn . Suppose M ⊂ Rn is a subspace that intersects Q at 0
only. Show that there is a hyperplane H, such that M ⊂ H, while H intersects Q at 0 only.
4. Bonus Problem, 3 Pt Total
Convex sets in infinite dimensional spaces can exhibit exotic properties, contradicting intuitions in finite
dimensional spaces. But for finite dimensional spaces, there could be some rather simple characterization of
convex sets.
1. Prove that any two norms ∥ · ∥1 , ∥ · ∥2 on Rn are equivalent in the sense that, there are constants a, b > 0,
such that
a∥x∥1 ≤ ∥x∥2 ≤ b∥x∥1 .
2. Let K ⊂ Rn be a bounded, closed convex subset. Prove that K must be homeomorphic to the closed unit
ball in some Rm with m < n.
