MAST30026 Problem Sets

Arun Ram
Department of Mathematics and Statistics
University of Melbourne
Parkville, VIC 3010 Australia
aram@unimelb.edu.au

Last update: 18 March 2014

Problem Set 1

  1. Check if the following functions are metrics on X.
    1. d(x,y)= |x2-y2| for x,y∈X=ℝ
    2. d(x,y)= |x2-y2| for x,y∈X=(-∞,0]
    3. d(x,y)= |arctan x-arctan y| for x,y∈X=ℝ
  2. (French railroad metric) Let X=ℝ2 and let d be the usual metric. Denote by 0=(0,0) and define d0(x,y)= { 0 if x=y; d(x,0)+ d(0,y) if x≠y. Verify that d0 is a metric on X. (Paris is at the origin 0.)
  3. Let X=ℝ2. For x=(x1,x2) and y=(y1,y2) define d(x,y)= { 1/2 if x1=y1  and x2≠y2  or if x1≠y1  and x2=y2; 1 if x1≠y1  and x2≠y2; 0 otherwise. Verify that d is a metric and that two congruent rectangles, one with base parallel to the x-axis and the other at 45∘ to the x-axis, have different “area” if d is used to measure the length of sides.
  4. Let (X,d) be a metric space. Consider the function f:[0,∞)→[0,∞) having the following properties:
    1. f is non-decreasing, i.e. f(a)≤f(b) if 0≤a<b;
    2. f(x)=0 if and only if x=0;
    3. f(a+b)≤f(a)+f(b), a,b∈[0,∞).
    If x,y∈X define df(x,y)=f(d(x,y)). Show that df is a metric and that the functions f(t)=kt where k>0, f(t)=tα where 0<α≤1, and f(t)=t1+t for t≥0 have properties (a)–(c).
  5. (p-adic metric) Let p be a prime number. Define the p-adic absolute value function |·|p on ℚ by setting |x|p=0 when x=0 and |x|p=p-k when x=pk·mn where m,n are nonzero integers which are not divisible by p. Show that for x,y∈ℚ, |x+y|p≤ max{|x|p,|y|p} and that d(x,y)=|x-y|p defines a metric on ℚ. In fact, d(x,z)≤max{d(x,y),d(y,z)}. If d satisfies this condition which is stronger than the triangle inequality then d is called an ultrametric.
  6. Let (Xi,di) be a metric space for 1≤i≤n and let X=∏i=1nXi. Define d(x,y) = [∑i=1ndi(xi,yi)2]1/2, d‾(x,y) = max { di(xi,yi)  | 1≤i≤n } , where x=(x1,…,xn) and y=(y1,…,yn)∈X. Verify that d and d‾ are metrics on X.
  7. Fix a positive integer n. Denote by 𝒫n the real vector space of all polynomials p(x)=anxn+an-1xn-1+⋯+a1x+a0 with real coefficients ai. For p(x)=anxn+an-1xn-1+⋯+a1x+a0∈𝒫n set ‖p‖=max { |a0|, |a1|,…, |an| } Verify that ‖·‖ is a norm on 𝒫n.
  8. Let (Xn,dn), n∈ℕ, be a sequence of metric spaces and let X=∏n∈ℕXn be the cartesian product of the Xn's. (The elements of X are of the form x=(x1,x2,…) with xn∈Xn.) For x,y∈X, define d(x,y)= ∑n=1∞ 12n dn(xn,yn) 1+dn(xn,yn) . Show that (X,d) is a metric space.
  9. Sketch the open ball B(0,1) in the metric space (ℝ3,di), where di is defined by d1(x,y) = |x1-y1|+ |x2-y2|+ |x3-y3| d2(x,y) = (x1-y1)2+ (x2-y2)2+ (x3-y3)2 d∞(x,y) = max { |x1-y1|, |x2-y2|, |x3-y3| } for x=(x1,x2,x3) and y=(y1,y2,y3)∈ℝ3.
  10. Set d(n,m)= |1n-1m| for n,m∈ℕ. Then d is a metric.
    1. Let P⊂ℕ be the set of positive even numbers. Find diam(P) and diam(ℕ\P) in (ℕ,d).
    2. For a fixed n∈ℕ find all elements of B(2n,12n) and B(n,12n).

Problem Set 2

  1. Let X=ℝ2. For x=(x1,x2) and y=(y1,y2)∈X define dM(x,y)= { |x2-y2| if x1=y1; |x1-y1|+ |x2|+ |y2| if x1≠y1. Also define dK(x,y)= { ‖x-y‖ if x=ty  for some t∈ℝ; ‖x‖+‖y‖ otherwise where ‖x‖=[∑i=1nxi2]1/2.
    (Can you give reasonable interpretations of the metrics dM and dK?)
    Study the convergence of the sequence xn in the spaces (X,dM) and (X,dK) if
    1. xn=(1n,nn+1);
    2. xn=(nn+1,nn+1);
    3. xn=(1n,n+1-n).
  2. Let {xn} and {yn} be sequences in a metric space (X,d) such that xn→n and yn→y as n→∞. Prove that d(xn,yn)→d(x,y) as n→∞.
  3. Metrics d and d‾ defined on X are called Lipschitz equivalent if there exist positive constants m,M such that m·d(x,y)≤ d‾(x,y) ≤M·d(x,y) for all x,y∈X.
    1. Show that if d and d‾ are Lipschitz equivalent, then they are equivalent. Give an example of X and two equivalent metrics on X which are not Lipschitz equivalent.
    2. For p≥1 and x,y∈ℝn, the lp metric is defined by dp(x,y)= [∑i=1n|xi-yi|p]1/p =‖x-y‖p. Show that if p,q≥1, then dp and dq are Lipschitz equivalent. (Hint: compare these with d∞(x,y)=max{|x1-y1|,…,|xn-yn|}.)
  4. Consider the set X=[-1,1] as a metric subspace of ℝ with the standard metric. Let
    1. A={x∈X | 1/2<|x|<2};
    2. B={x∈X | 1/2<|x|≤2};
    3. C={x∈ℝ | 1/2≤|x|<1};
    4. D={x∈ℝ | 1/2≤|x|≤1};
    5. E={x∈ℝ | 0<|x|≤1 and 1/x∉ℤ}.
    Classify the sets in (a)–(e) as open/closed in X and ℝ.
  5. Consider ℝ2 with the standard metric. Let
    1. A={(x,y) | -1<x≤1 and -1<y<1};
    2. B={(x,y) | xy=0};
    3. C={(x,y) | x∈ℚ,y∈ℝ};
    4. D={(x,y) | -1<x<1 and y=0};
    5. E=⋃n=1∞{(x,y) | x=1/n and |y|≤n}.
    Sketch (if possible) and classify the sets in (a)–(e) as open/closed/neither in ℝ2.
  6. Find the interior, the closure and the boundary of each of the following subsets of ℝ2 with the standard metric:
    1. A={(x,y) | x>0 and y≠0};
    2. B={(x,y) | x∈ℕ,y∈ℝ};
    3. C=A∪B;
    4. D={(x,y) | x is rational};
    5. E={(x,y) | x≠0 and y≤1/x}.
  7. Let A be a subset of a metric space X. Is the interior of A equal to the interior of the closure of A? Is the closure of the interior of A equal to the closure of A itself?
  8. Consider a collection {Ai}i∈I of subsets of a metric space X. Show that ⋃i∈IAi∘⊂ (⋃i∈IAi)∘ ⋂i∈IAi‾⊂ ⋂i∈IAi‾ (⋂i∈IAi)∘⊂ ⋂i∈IAi∘ ⋃i∈IAi‾⊂ ⋃i∈IAi‾
  9. Let (X,d) be a metric space. Show that if A⊂X, then
    1. A‾=A∪∂A.
    2. ∂A=A‾\A∘ and A∘=A\∂A.
    3. A is closed if and only if ∂A=A\A∘.
    4. A is open if and only if ∂A=A‾\A.
  10. Let X and Y be metric spaces and A,B non-empty subsets of X and Y, respectively. Prove that
    1. If A×B is an open subset of X×Y, then A and B are open in X and Y, respectively.
    2. If A×B is an closed subset of X×Y, then A and B are closed in X and Y, respectively.

Problem Set 3

  1. Let d and d′ be equivalent metrics on X. Show that
    1. A⊂X is closed in (X,d) if and only if A is closed in (X,d′);
    2. A⊂X is open in (X,d) if and only if A is open in (X,d′).
  2. Show that if A⊂X then diam A=diam A‾. Does diam A=diam A0?
  3. Let (X,dX) and (Y,dY) be metric spaces and A,B are dense subsets of X and Y, respectively. Show that A×B is dense in X×Y.
  4. Let C be the circle in ℝ2 with the centre at (0,1/2) and radius 1/2. Let X=C\{(0,1)}. Define the function f:ℝ→X by defining f(t) to be the point at which the line segment from (t,0) to (0,1) intersects X.
    1. Show that f:ℝ→X and f-1:X→ℝ are continuous.
    2. Define for s,t∈ℝ ρ(s,t)= ‖f(s)-f(t)‖ where ‖·‖ is the standard norm in ℝ2. Show that ρ defines a metric on ℝ which is equivalent to the standard metric on ℝ.
  5. Let X=C[0,1]. Let F:X→ℝ be defined by F(f)=f(0). Moreover, let d∞(f,g)=sup{|f(x)-g(x)|:x∈[0,1]} and d1(f,g)=∫01|f(x)-g(x)|dx.
    Is F continuous when X is equipped with (a) the metric d∞, (b) the metric d1?
  6. Let (X,dX) and (Y,dY) be metric spaces. Show that f:X→Y is continuous if and only if
    1. f(A‾)⊂f(A)‾ for all subsets A of X, or
    2. f-1(B)‾⊂f-1(B‾) for all subsets B of Y.
  7. Let (X,d) be a metric space and let a be a fixed point of X. Show that |d(x,a)-d(y,a)| ≤d(x,y) for all x,y∈X. Conclude that the function f:X→ℝ defined by f(x)=d(x,a) is uniformly continuous.
  8. Which of the following functions are uniformly continuous?
    1. f(x)=sin x on [0,∞)
    2. g(x)=11-x on (0,1)
    3. h(x)=x on [0,∞)
    4. k(x)=sin(1/x), on (0,1)
  9. Which of the following sequences of functions converge uniformly on the interval [0,1]?
    1. fn(x)=nx2(1-x)n
    2. fn(x)=n2x(1-x2)n
    3. fn(x)=n2x3e-nx2
  10. Suppose that A is a dense subset of a metric space (X,d) and f:A→ℝ is uniformly continuous. Show that there exists exactly one continuous function g:X→ℝ satisfying g(x)=f(x) for x∈A.
    (Hint: You may need to use the completeness of ℝ.)

Problem set 4

  1. Suppose that {xn} and {yn} are Cauchy sequences in a metric space (X,d). Prove that the sequence of real numbers {d(xn,yn)} converges.
  2. Suppose that {xn} is a sequence in a metric space (X,d) such that d(xn,xn+1)≤2-n for all n∈ℕ. Prove that {xn} is a Cauchy sequence.
  3. Decide if the following metric spaces are complete:
    1. ((0,∞),d), where d(x,y)=|x2-y2| for x,y∈(0,∞).
    2. ((-π/2,π/2),d), where d(x,y)=|tan x-tan y| for x,y∈(-π/2,π/2).
  4. Let X=(0,1] be equipped with the usual metric d(x,y)=|x-y|. Show that (X,d) is not complete. Let d∼(x,y)=|1x-1y| for x,y∈X. Show that d∼ is a metric on X that is equivalent to d, and that (X,d∼) is complete.
  5. Suppose that (X,d) and (Y,d∼) are metric spaces and that f:X→Y is a bijection such that both f and f-1 are uniformly continuous. Show that (X,d) is complete if and only if (Y,d∼) is complete.
  6. [Cantor’s Intersection Theorem] Let (X,d) be a metric space and let {Fn} be a “decreasing” sequence of non-empty subsets of X satisfying Fn+1⊆Fn for all n.
    1. Prove that if
      1. (X,d) is complete,
      2. each Fn is closed,
      3. diam(Fn)→0,
      then ⋂n∈ℕFn consists of exactly one point.
    2. Show that, if any of (i)-(iii) is omitted, then ⋂n∈ℕFn may be empty.
    3. Conversely, prove that if for every decreasing sequence {Fn} of non-empty subsets satisfying (ii) and (iii), the intersection ⋂n∈ℕFn is non-empty, then X is complete.
  7. Let (X,d) be a complete metric space and let f:X→(0,∞) be a continuous function. Prove that there exists a point x* such that f(y)≤2f(x*) for all y∈B(x*,1f(x*)).
    (Hint: Arguing by contradiction show that there exists a sequence {xn} with the following properties: f(x1)>0, f(xn+1)>2f(xn) for all n≥1 and d(xn+1,xn)≤1f(xn). Then show that {xn} is Cauchy.)

Problem Set 5

  1. Let {fk} be a sequence of linear maps fk:ℝn→ℝm which are not identically zero, that is, for every k∈ℕ there is x=xk such that fk(x)≠0. Show that there is x (not depending on k) such that fk(x)≠0 for all k∈ℕ.
  2. Let {fn} be a sequence of continuous functions fn:ℝ→ℝ having the property that {fn(x)} is unbounded for all x∈ℚ. Prove that there is at least one x∈ℚc such that {fn(x)} is unbounded.
  3. Let (X,d) be a complete metric space and let (Y,d∼) be a metric space. Let {fn} be a sequence of continuous functions from X to Y such that {fn(x)} converges for every x∈X. Prove that for every ε>0 there exist k∈ℕ and a non-empty open subset U of X such that d∼(fn(x),fm(x))<ε for all x∈U and all n,m≥k.
  4. Which of the following maps are contractions?
    1. f:ℝ→ℝ, f(x)=e-x;
    2. f:[0,∞)→[0,∞), f(x)=e-x;
    3. f:[0,∞)→[0,∞), f(x)=e-ex;
    4. f:ℝ→ℝ, f(x)=cos x;
    5. f:ℝ→ℝ, f(x)=cos(cos x).
  5. Consider the map f:ℝ2→ℝ2 given by f(x,y)=110 (8x+8y,x+y),  (x,y)∈ℝ2. Recall metrics d1((x1,y1),(x2,y2))=|x1-x2|+|y1-y2|, d2((x1,y1),(x2,y2))=[|x1-x2|2+|y1-y2|2]1/2 and d∞((x1,y1),x2,y2)=max{|x1-x2|,|y1-y2|}. Is f a contraction with respect to d1? d2? d∞?
    1. Consider X=(0,a] with the usual metric and f(x)=x2 for x∈X. Find values of a for which f is a contraction and show that f:X→X does not have a fixed point.
    2. Consider X=[1,∞) with the usual metric and let f(x)=x+1x for x∈X. Show that f:X→X and d(f(x),f(y))<d(x,y) for all x≠y, but f does not have a fixed point.
      Reconcile (a) and (b) with Banach fixed point theorem.
  6. 7. Let (X,d) be a complete metric space and f:X→X be a function such that d(f(x),f(y)) ≤αd(x,y) for all x,y∈B‾(x0,r0), where 0<α<1 and d(x0,f(x0))≤(1-α)·r0. Prove that f has a unique fixed point p∈B‾(x0,r0).
    1. Show that there is exactly one continuous function f:[0,1]→ℝ which satisfies the equation [f(x)]3-ex [f(x)]2+ f(x)2=ex. (Hint: rewrite the equation as f(x)=ex+12f(x)1+f(x)2.)
    2. Consider C[0,a] with a<1 and T:C[0,a]→C[0,a] given by (Tf)(t)=sin t+ ∫0tf(s)ds, t ∈[0,a]. Show that T is a contraction. What is the fixed point of T?
    3. Find all f∈C[0,π] which satisfy the equation 3f(t)=∫0tsin (t-s)f(s)ds.
    4. Let g∈C[0,1]. Show that there exists exactly one f∈C[0,1] which solves the equation f(x)+∫01 ex-y-1f(y) dy=g(x), for all  x∈[0,1]. (Hint: Consider the metric d(f,h)=sup{e-x|f(x)-h(x)|:x∈[0,1]}.)

Problem Set 6

  1. On ℝ consider the metrics: d1(x,y) = |arctan x-arctan y|, d2(x,y) = |x3-y3|. With which of these metrics is ℝ complete? If (ℝ,di) is not complete find its completion.
  2. Which of the following subsets of ℝ and ℝ2 are compact? (ℝ and ℝ2 are considered with the usual metrics).
    1. A=ℚ∩[0,1]
    2. B={(x,y)∈ℝ2:x2+y2=1}
    3. C={(x,y)∈ℝ2:x2+y2<1}
    4. D={(x,y):|x|+|y|≤1}
    5. E={(x,y):x≥1 and 0≤y≤1/x}
  3. Prove that if A1,…,Ak are compact subsets of a metric space (X,d), then ⋃i=1kAi is compact.
  4. Prove that if Ai is a compact subset of the metric space (Xi,di) for i=1,…,k, then A1×⋯×Ak is a compact subset of X=X1×⋯×Xk with the product metric d.
  5. Let A be a non-empty compact subset of a metric space (X,d). Prove:
    1. If x∈X, then there exists a∈A such that d(x,A)=d(x,A);
    2. If A⊂U and U is open, then there is ε>0 such that {x∈X:d(x,A)<ε}⊂U.
    3. If B is closed and A∩B=∅, then d(A,B)>0.
    Hint: Recall that (x,y)↦d(x,y) is continuous from X×X→[0,∞).
  6. Let f:X→ℝ. Call a function f upper semicontinuous, abbreviated u.s.c., if for every r∈ℝ, {x∈X | f(x)<r} is open. Similarly, f is lower semicontinuous, abbreviated l.s.c., if for every r∈ℝ, {x∈X | f(x)>r} is open. Assume that X is compact. Show that every u.s.c. function assumes a maximum value and every l.s.c. function assumes a minimum value.
  7. Call a map f:X→X a weak contraction if d(f(x),f(y))<d(x,y) for all x≠y. Prove that if X is compact and f is a weak contraction, then f has a unique fixed point.

The next problem gives a different construction of the completion of a metric space (X,d).

An equivalence relation on a set X is a relation ∼ having the following three properties:

  1. (Reflexivity) x∼x for every x∈X.
  2. (Symmetry) If x∼y, then y∼x.
  3. (Transitivity) If x∼y and y∼z, then x∼z.
The equivalence class determined by x, and denoted by [x], is defined by [x]={y∈X:y∼x}. We have [x]=[y] if and only if x∼y, and X is a disjoint union of these equivalence classes.

  1. Let (X,d) be a metric space and let X* be the set of Cauchy sequences x={xn} in (X,d). Define a relation ∼ in X* by declaring x={xn}∼y={yn} to mean d(xn,yn)→0.
    1. Show that ∼ is an equivalence relation.
      Denote by [x] the equivalence class of x∈X*, and let X∼ denote the set of these equivalence classes.
    2. Show that if x={xn} and y={yn}∈X*, then limn→∞d(xn,yn) exists. Show that if x′={xn′}∈[x] and y′={yn′}∈[y], then limn→∞d (xn,yn)= limn→∞d (xn′,yn′). For [x],[y]∈X∼, define D([x],[y]) =limn→∞d (xn,yn). Note that the definition of D is unambiguous in view of the above equality.
    3. Show that (X∼,D) is a complete metric space.
      Hint: Let [xn] be Cauchy in (X∼,D). Then xn={x1n,x2n,x3n,…} is Cauchy in (X,d). So for every n∈ℕ, there exists kn∈ℕ such that d(xmn,xknn) <1/n for all m≥kn. Set xn={x1n,x2n,x3n,…}. Then show that x is Cauchy in (X,d) and D([xn],[x])→0.
    4. If x∈X, let φ(x) be the equivalence class of the constant sequence x=(x,x,x,…). That is, φ(x)=[x]=[{x,x,x,…}]. Show that φ:X→φ(X) is an isometry.
    5. Show that φ(X) is dense in (X∼,D).
      Hint: Let [x]∈X∼ with x={x1,x2,x3,…}. Denote by xn the constant sequence {xn,xn,xn,…} and show that D([xn],[x])→0.

Problem Set 7

  1. Consider the following spaces:
    1. ℝ with the metric d1(x,y)= |x-y|1+|x-y|;
    2. ℝ with the metric d2(x,y)= |arctan x-arctan y|;
    3. ℝ with the metric d3(x,y)=0 if x=y and d(x,y)=1 if x≠y.
    Is (ℝ,di) compact?
  2. Use the Heine-Borel property to prove that if f:X→Y is a continuous mapping between metric spaces and X is compact then f is uniformly continuous.
  3. A family {Fi}i∈I is said to have the finite intersection property if for every finite subset J of I, ⋂i∈JFi≠∅. Show that X is compact if and only if for every family {Fi}i∈I of closed subsets of X having the finite intersection property, the intersection ⋂i∈IFi≠∅.
  4. Consider C[0,1] with the usual d∞ metric. Let A= { f∈C[0,1]:0=f (0)≤f(t)≤f (1)=1 for all t∈ [0,1] } . Show that there is no finite 1/2-net for A.
  5. Show that if A⊂X is totally bounded, then A‾ is also totally bounded.
  6. Show that a metric space (X,d) is totally bounded if and only if every sequence {xn}⊆X contains a Cauchy subsequence.
  7. Let X be a totally bounded metric space and Y a metric space. Assume that f:X→Y is a bijection. Show that if f and f-1 are uniformly continuous, then Y is totally bounded.
  8. [Lebesgue number lemma] Let (X,d) be a compact metric space and let {Ui}i∈I be an open covering of X. Prove that there exists δ>0 such that for every subset A⊂X with diam(A)<δ there exists i∈I such that A⊂Ui.
    (δ is called a “Lebesgue number” for the covering.)
  9. Let (X,d) be a compact metric space. Assume that f:X→X preserves distance, that is, d(f(x),f(y)) =d(x,y) for every x,y∈X. Show that f is a bijection.
    Hint: Assume that f(X)≠X. So there exists a∈X\f(X). Since f is continuous and X is compact, f(X) is compact. So d(a,f(X))=r>0. Consider a sequence xn=fn(a).

Problem Set 8

  1. Which of the following sets X are connected in ℝ2?
    1. Let H={(x,y)∈ℝ2 | xy=1 and x,y>0}, L={(x,0) | x∈ℝ}, and X=H∪L;
    2. Let Cn={(x,y)∈ℝ2 | (x-1/n)2+y2=1/n2} for n∈ℤ, and X=⋃n∈ℕCn.
  2. Show that if A is a connected subspace of a topological space (X,𝒯) and if A⊂B⊂A‾, then B is connected.
  3. If A and B are connected subsets of a topological space (X,𝒯) such that A‾∩B≠∅, then A∪B is connected.
  4. A point p∈X is called a cut point if X\{p} is disconnected. Show that the property of having a cut point is a topological property. (A property of a topological space is a topological property if it is preserved under homeomorphisms.)
  5. Show that no two of the intervals (a,b),(a,b], and [a,b] are homeomorphic.
  6. Show that ℝ and ℝ2 are not homeomorphic (where ℝ and ℝ2 are equipped with the usual topologies).
  7. Let S1={(x,y)∈ℝ2 | x2+y2=1} be the unit circle in ℝ2, and let f:S1→ℝ be a continuous function. Show that there exists x∈S1 such that f(x)=f(-x). [Hint: consider the function g:S1→ℝ where g(x)=f(x)-f(-x).]
  8. Let A be a countable set. Show that ℝ2\A is path connected.
  9. Show that if A is an open connected subset of ℝn, then A is path connected. [Hint: Fix a point x0∈A and consider the set U of all x∈A which can be joined to x0 by a path in A. Show that U and A\U are open.]
  10. A metric space (X,dX) is called a chain connected if for every pair x,y of points in X and every ε>0, there are finitely many points x=x0,x1,x2,…,xn=y such that dX(xi+1,xi)<ε for i=0,1,…,n-1. Prove that a compact, chain connected metric space is connected.

Problem Set 9

  1. Show that ℝn becomes a real inner product space with ⟨X,Y⟩=XTAY, where X,Y∈ℝn are column vectors and A is a real symmetric matrix with positive eigenvalues. Similarly show that ℂn is a complex inner product space with ⟨X,Y⟩=XTBY, where B is a Hermitian matrix with positive eigenvalues. (Recall that a matrix B is Hermitian if B‾T=B, ie B is equal to the result of taking the complex conjugate transpose of B).
  2. Show that ‖x‖=sup|⟨x,y⟩|‖y‖, over all y≠0 in any inner product space.
  3. Let H be the Hilbert space L2[-1,1]. Show that Gram Schmidt applied to the total set {1,t,t2,t3,…} yields an orthonormal basis which is the sequence of Legendre polynomials given by Lk(t)=ck dkdtk (t2-1)k, k=1,2,3,… where the ck are determined by requiring the polynomials to have unit length. In particular, show that the polynomials are orthogonal for any choice of ck. (You don’t need to compute the ck).
  4. Let W be a subspace of a Hilbert space H which admits an orthogonal projection P. Show;
    1. P2=P
    2. dist(x,W)=‖x-Px‖, ie Px is the closest point to x in W.
  5. Let A,B be subsets of a Hilbert space H. If A⊥ is the orthogonal complement of A, defined by A⊥={x∈H:⟨x,a⟩=0,a∈A} then prove;
    1. A⊥ is a closed subspace of H
    2. A∩A⊥⊂{0}
    3. A⊂B⇒A⊥⊃B⊥
    4. A⊂A⊥⊥.
    5. If W is a subspace of H then W is closed if and only if W=W⊥⊥.
  6. Let S={e1,e2,…} be a countable orthonormal basis for a separable Hilbert space H. Prove;
    1. x=∑n⟨x,en⟩en (Fourier series)
    2. ‖x‖2=∑n‖x,en‖2 (Parseval's identity for norms)
    3. ⟨x,y⟩=∑n⟨x,en⟩⟨y,en⟩‾ (Parseval's identiy for inner products)
  7. If W,V are closed subspaces of a Hilbert space H and W⊥V then show that W+V is closed.
  8. Let S be a subset of a Hilbert space H satisfying S⊥={0}. Show that S is a total set in H, i.e ⟨S⟩‾=H, ie the closure of the span of S is H.
  9. Let S be an arbitrary set. By l2(S) we mean the set of all functions f:S→ℂ such that f(s)≠0 for countably many s∈S and such that the series ∑{s∈S}|f(s)|2 converges. Define ⟨f,g⟩=∑{s∈S}⟨f(s),g(s)⟩ for f,g∈l2(S). Prove that;
    1. l2(S) is a Hilbert space.
    2. l2=l2(ℕ)
    3. Every Hilbert space with an orthonormal basis S is isometric to l2(S).
    (Hint: define functions fe:S→ℂ by fe(e)=1, fe(e′)=0 for e,e′∈S, e≠e′. Show that S′={fe:e∈S} is an orthonormal basis for l2(S) and the bijection e→fe extends to an isometry H→l2(S).)

Problem Set 10

  1. Let R and L be the left and right shift operators in the normed space lp. So R(a1,a2,a3,…)= (0,a1,a2,…), L(a1,a2,a3,…)= (a2,a3,…) for all (a1,a2,…)∈lp.
    1. Show that R,L are bounded linear operators.
    2. Show that R is injective but not surjective and L is surjective but not injective.
    3. Show that LR=I but RL≠I
    4. Show that ‖Ln(x)‖→0 for all x∈lp but ‖Ln‖↛0.
    5. Find the norms ‖L‖,‖R‖.
  2. For the case p=2 find the adjoints of the shift operators R,L:l2→l2.
  3. Let a be a fixed element of a Hilbert space H. Prove that the mapping f(x)=⟨x,a⟩ is a bounded linear functional with ‖f‖=‖a‖.
  4. Prove the following facts about adjoints of bounded linear operators on Hilbert spaces.
    1. (T+S)*=T*+S*
    2. (TS)*=S*T*
    3. (λT)*=λ‾T*
    4. ‖T*T‖=‖T‖2
  5. Let Sn be a sequence of self adjoint operators on a Hilbert space H which converge pointwise to a bounded linear operator S. Show that S is self adjoint.
  6. If T is a positive operator on a Hilbert space H prove that Tn is positive for all n≥1.
  7. Prove that if T is a positive operator then every eigenvalue of T is non-negative.
  8. Let P be an orthogonal projection on a Hilbert space H. Prove that P is self adjoint, positive and I-P is positive.
  9. Let T:ℂn→ℂm be a linear transformation with matrix [ajk] relative to the standard basis, so that the matrix is m×n. Suppose the norms in ℂn,ℂm are both the supremum norm. Show that ‖T‖=maxj∑k|ajk|.
  10. Let T:ℂn→ℂm be a linear transformation with matrix [ajk] relative to the standard basis, so that the matrix is m×n. Suppose the norms in ℂn,ℂm are both the l1 norms. Show that ‖T‖=maxk∑k|ajk|.

Problem Set 11

  1. Let T be a bounded self adjoint compact operator on a Hilbert space H. If λ is a non zero complex number so that λI-T is a one-to-one mapping, prove that λI-T is onto and has a bounded inverse. (Hint: Use the spectral theorem).
  2. Let T be a bounded self adjoint compact operator on a Hilbert space H. If λ is a non zero complex number so that λI-T is an onto mapping, prove that λI-T is one-to-one and has a bounded inverse. (Hint: Show that if N is the null space for λI-T and R is the closure of the range of λI-T then N=R⊥.)
  3. For the final example in the notes, check that the functions sn(t) are indeed eigenvectors for the Fredholm integral operator T associated to the Green’s function G with eigenvalues n2.
  4. Check that the functions sn(t) form an orthonormal set in L2[0,π].
  5. Check that the eigenvectors for L are the solutions to the Sturm Liouville system in the last example y″+λy=0 on [0,π] with the boundary conditions y(0)=y(π)=0.

page history