Numerical Range of Operators on

Hilbert Space
Debarati Bhattacharya
Student ID- 12310070

Supervised By
Dr. Arnab Patra

Indian Institute of Technology Bhilai

Winter Semester (2023-24) Research Report

June 4, 2024
1 Activities in First Semester
In the semester I, I’ve studied Introduction to Spectral T heory, F unctional Analysis,
Operator T heory II and Additive N umber T heory. Course name, course code and
obtained grades are given in the following table.

Grades of First Semester

Course Name Course Code Grade

Introduction to Spectral Theory MAL600 A−
Functional Analysis MAL502 B
Operator Theory II MAL602 C
Additive Number Theory MAL511 B

1
2 Activities in Second semester

In this semester I have opted the course T opology (MAL500) which is very much helpful
for my research work and I obtained A+ .
Beside this I have also studied Real Analysis, M easure T heory, Banach Algebra, C ∗ Algebra
and some basics of N umerical Ranges of Operators on Hilbert Space .

2.1 Real Analysis and Measure Theory

Some well-known theorems and results are as follows:

Theorem 1 (Lebesgue’s Monotone Convergence Theorem). Let {fn } be a sequence

of measurable functions on X, and suppose that
(a) 0 ≤ f 1 (x) ≤ f 2 (x) ≤ ... ≤ ∞ for every x ∈ X.
(b) f n (x) → f (x) as n → ∞, R for everyRx ∈ X.
Then f is measurable and X f n dµ → X f dµ as n → ∞.

LemmaR 1 (Fatou’s Lemma). If f n : X →R [0, ∞] is measurable, for each positive integer

n, then X (lim inf n→∞ f n )dµ ≤ lim inf n→∞ X f n dµ.

Theorem 2 (Lebesgue’s Dominated Convergence Theorem). Suppose {fn } be a

sequence of complex measurable functions on X such that f (x) = limn→∞ f n (x) exists for
every x ∈ X. If there is a function
R g ∈ L1 (µ) such that |f n (x)|R≤ g(x) (nR= 1, 2, 3, ..., x ∈
1
X), then f ∈ L (µ), limn→∞ X |f n − f |dµ = 0, and limn→∞ X f n dµ = X f dµ.

Lemma 2 (Urysohn’s Lemma). Suppose X is a locally compact Hausdorff space, V

is open in X, K ⊂ V and K is compact. Then there exists an f ∈ C c (X), such that
0 ≤ f (x) ≤ 1 for all x ∈ X and that f (x) = 1 for all x ∈ K and Supp(f ) ⊂ V .

Theorem 3 (The Riesz Representation Theorem). Let X be a locally compact

Hausdorff space and let Λ be a positive linear functional on C c (X). Then there exists a
σ-algebra ℳ in X which contains all Borel sets in X and there exists a unique positive
measure µ on ℳ which represents Λ in the sense that

R
(i) Λf = X
f dµ for every f ∈ C C (X), and which has the following additional proper-
ties:

(ii) µ(K) < ∞ for every compact set K ⊂ X.

(iii) For every E ∈ ℳ, we have µ(E) = inf {µ(V ) : E ⊂ V, V open}.

(iv) The relation µ(E) = sup{µ(K) : K ⊂ E, Kcompact}.

(v) If E ∈ ℳ, A ⊂ E and µ (E) = 0, then A ∈ ℳ.

Theorem 4. If A ⊂ R1 and every subset of A is Lebesgue measurable then m(A) = 0.

Corollary 1. Every set of positive measure has nonmeasurable subsets.

2
Theorem 5 (Lusin’s Theorem). Suppose that f is a complex measurable function on
X, µ(A) < ∞, f (x) = 0 if x ∈
/ A and ϵ > 0.Then there exists a g ∈ Cc (X) such that
µ({x : f (x) ̸= g(x)}) < ϵ.
Furthermore, we may arrange it so that supx∈X |g(x)| ≤ supx∈X |f (x)|.

Theorem 6 (The Vitali-Caratheodory Theorem). Suppose f ∈ L1 (µ),f is real-

valued, and ϵ > 0. Then there exist functions u and v on X such that u ≤ f ≤ v, u is
upperR semicontinuous and bounded above, v is lower semicontinuous and bounded below,
and X (v − u)dµ < ϵ.

Theorem 7 (Jensen’s Ineqality). Let µ be a positive measure on a σ-algebra ℳ in a

1
R functionR in L (µ), if a < f (x) < b for all x ∈ Ω,
set Ω, so that µ(Ω) = 1. If f is a real
and if ϕ is convex on (a, b), then ϕ( Ω f dµ) ≤ Ω (ϕ ◦ f )dµ.
Note: The cases a = −∞ and b = ∞ are not excluded and in this case ϕ ◦ f is not in
L1 (µ), it exists in the extended sense and its value is +∞.

Definition 1. If 0 < p < ∞ and if f is a complex measurable function on X, define

R 1/p
∥f ∥p = { X |f |p dµ} , where X is an arbitrary measure space with a positive measure µ
and let Lp (µ) consist of all f for which ∥f ∥p < ∞.

Theorem 8. Lp (µ) is a complete metric space, for 1 ≤ p ≤ ∞ and for every positive
measure µ.

Theorem 9. Let S be the class of all complex, measurable, simple functions on X such
that µ({x : s(x) ̸= 0}) < ∞. If 1 ≤ p < ∞, then S is dense in Lp (µ).

Theorem 10. For 1 ≤ p < ∞, Cc (X) is dense in Lp (µ).

3
2.2 Banach Algebra
Definition 2 (Algebra). An algebra over a field F is a vector space 𝒜 over F that also
has a multiplication defined on it that makes 𝒜 into a ring such that if α ∈ F and a, b ∈
𝒜, α(ab) = (αa)b = a(αb).

Definition 3 ( Banach algebra). A Banach algebra is an algebra 𝒜 over F that has

a norm ∥.∥ relative to which 𝒜 is a Banach space and such that for all a, b in 𝒜,
∥ab∥ ≤ ∥a∥∥b∥

Definition 4 (Spectral radius). If 𝒜 is a Banach algebra with identity and a ∈ 𝒜, the

spectralradius of a, r(a), is defined by
r(a) = sup{|α| : α ∈ σ(a)}

Proposition 1. If 𝒜 is a Banach algebra with identity and a ∈ 𝒜, lim∥an ∥1/n exists and
r(a) = ∥an ∥1/n

Theorem 11 (Gelfand-Mazur theorem). There is a canonical bijection between the

maximal ideals in a commutative unital Banach algebra 𝒜 and its characters given by
associating to each character its kernel.

Definition 5 (Gelfand spectram). If 𝒜 is commutative Banach algebra with unity,

Ω(𝒜) = { τ : 𝒜 → C : τ is nonzero, multiplicative} is called the Gelf and spectram of
𝒜.

Theorem 12. If 𝒜 is a Banach algebra with identity and a ∈ 𝒜,

σ(a) = {h(a) : h ∈ Ω(𝒜)}.

Theorem 13. Ω(𝒜) is ω ∗ compact Hausdorff space.

Definition 6. If 𝒜 is a Banach algebra with identity and a ∈ 𝒜, the Gelf and transf ormation
of 𝒜 is defined by
â : Ω(𝒜) → C such that â(τ ) = τ (a).

Theorem 14. If 𝒜 is commutative Banach algebra with unity,

(i) For a ∈ 𝒜, â is a continuous function, i.e. â ∈ C(Ω(𝒜))

(ii) ∥â∥ = r(a)

(iii) The map Γ𝒜 : 𝒜 → C(Ω(𝒜)) such that Γ𝒜 (a) = â is a homomorphism of Banach

algebra and Γ𝒜 is continuous.

4
2.3 C∗ -Algebra
Definition 7 (Banach ∗ −algebra). A Banach algebra 𝒜 is said to be a Banach ∗
−
algebra if there is an operation ∗ : a → a∗ satisfying:

1. The map ∗ : a → a∗ is conjugate linear, i.e., (λa + µb)= λa+ µb∗ ∀a, b ∈ 𝒜 and
∀λ, µ ∈ C

2. (a∗ )∗ = a

3. (ab)∗ = b∗ a∗

4. ∥a∗ ∥ = ∥a∥

Definition 8 (C ∗ −Algebra). A Banach ∗ - algebra is said to be a ∗ − algebra if it satisfies

the C ∗ condition
∥a∗ a∥ = ∥a∥2 ∀a ∈ 𝒜.

Theorem 15 (Gelfand-Neimark). Suppose that 𝒜 is a commutative unital Banach∗ −

algebra. The Gelfand transform b: 𝒜 → C(Ω(𝒜)) is an isometric ∗ -isomorphism if and
only if 𝒜 is a C ∗ -algebra.

Theorem 16. Let a be a normal element of a unital C ∗ -algebra 𝒜. Then there is a

unique isomorphism π : C(σ(a)) → C ∗ {a, 1} such that π(p) = p(a) for every polynomial
p.

Theorem 17 (Spectral mapping theorem for continuous functions). Let a be a

normal element of a unital C ∗ -algebra 𝒜 and f ∈ C(σ(a)). Then
f (σ(a)) = σ(f (a)).

Definition 9 (Positive element). Let 𝒜 be a unital C ∗ -algebra. Then an element a ∈ 𝒜

is said to be positive if a = bb∗ for sobe b ∈ 𝒜.

Theorem 18 (Existance of positive square root). Let a be a self-adjoint element of

a unital C ∗ -algebra 𝒜. Then if σ(a) ⊆ R+ , there exists b = b∗ with σ(b) ⊆ R+ such that
a = b2 .

Theorem 19. For any element a in a unital C ∗ -algebra 𝒜

a∗ a ≤ ∥a∥2 .1
i.e., ∥a∥2 .1 − a∗ a is a positive element of the C ∗ -algebra.

Definition 10 (Positive linear functional). Let 𝒜 be a unital C ∗ -algebra. Then a

linear functional ϕ : 𝒜 → C is said to be positive linear functional if ϕ(a) ≥ 0 for a ≥ 0.

Definition 11 (State). A positive linear functional ϕ is said to be a state if ∥ϕ∥ = 1.

Lemma 3. Let 𝒜 be a unital C ∗ -algebra and ϕ : 𝒜 → C be a linear functional. Then ϕ

is positive if f the norm of ϕ is attained at identity, i.e., ∥ϕ∥ = ϕ(1).
In particular, every positive linear functional is bounded and it is a state if f ∥ϕ∥ =
ϕ(1) = 1.

Definition 12. A representation of a C ∗ -algebra 𝒜 on a Hilbert space H is a ∗-homomorphism

π : 𝒜 → B(H) where B(H) is the C ∗ -algebra of all bounded operators on H.

5
Definition 13. A representation of a C ∗ -algebra 𝒜 on a Hilbert space H is a ∗-homomorphism
π : 𝒜 → B(H) where B(H) is the C ∗ -algebra of all bounded operators on H.

Definition 14. A representation π : 𝒜 → B(H) is said to be cyclic if there exists a

vector v in H such that π(a)v is dense in H.

Theorem 20 (GNS Construction for States). Let 𝒜 be a unital C ∗ -algebra and

let ϕ be a state on 𝒜. Then there exists a triple (H, π, u) where H is a Hilbert space,
π : 𝒜 → B(H) is a unital ∗-homomorphism and u ∈ H is a unit vector such that
ϕ(a) = ⟨u, π(a)u⟩ for all a ∈ 𝒜. The triple (H, π, u) satisfies ’minimality condition’ that
is H = Span{π(a)u : a ∈ 𝒜}. Moreover, such a triple is unique upto unitary equivalence,
that is given another such triple (H ′ , π ′ , u′ ), there exists a unitary U : H → H ′ such that
U u = u′ and π ′ (a) = U π(a)U ∗ ∀a ∈ 𝒜.

Theorem 21 (Decomposition to cyclic representation). Let π be a representation

of a C ∗ -algebra 𝒜 on a HilbertL
space H. Then π
Ldecomposes as a direct sum of a family
of cyclic representations: π = i∈I πi on H = i∈I Hi , where each πi is cyclic.

Theorem 22 (Gelfand, Naimark and Segal). Let 𝒜 be a unital C ∗ -algebra. Then

there exists a pair (H, π) where H is a complex Hilbert space and π : 𝒜 → B(H) is an
injective ∗-homomorphism.

Corollary 2. If the C ∗ -algebra 𝒜 is separable then each of the GNS Hilbert spaces Ha
are separable.

Theorem 23 (Polar Decomposition Theorem). Let a be an invertible element in

a unital C ∗ -algebra 𝒜. Then there exists a unique unitary U such that a = U |a| where
|a| = (a∗ a)1/2 .

Theorem 24 (Kadison and Pederson). Let 𝒜 be a unital C ∗ -algebra.Let a ∈ 𝒜 be in

the unit ball of 𝒜 such that ∥a∥ < 1 − 2/n, for some n ∈ N with n ≥ 3. Then there exists
n unitaries in 𝒜 such that a = (U1 + U2 + ...Un )/n.

Theorem 25 (Russo Dye Theorem). Let 𝒜 be a unital C ∗ -algebra. Then its closed
unit ball is the closed convex hull of its unitary elements.