ConceptComplete

Bounded Linear Operators - Key Properties

The adjoint operator is a fundamental construction that extends the notion of transpose from finite-dimensional linear algebra to infinite-dimensional Banach and Hilbert spaces.

DefinitionAdjoint Operator (Banach Space)

Let X,YX, Y be Banach spaces and T∈B(X,Y)T \in \mathcal{B}(X, Y). The adjoint T∗:Y∗→X∗T^* : Y^* \to X^* is defined by ⟨T∗(ϕ),x⟩=⟨ϕ,T(x)⟩\langle T^*(\phi), x \rangle = \langle \phi, T(x) \rangle for all ϕ∈Y∗\phi \in Y^* and x∈Xx \in X.

The adjoint T∗T^* is a bounded linear operator with ∥T∗∥=∥T∥\|T^*\| = \|T\|.

DefinitionAdjoint Operator (Hilbert Space)

Let H1,H2H_1, H_2 be Hilbert spaces and T∈B(H1,H2)T \in \mathcal{B}(H_1, H_2). The Hilbert space adjoint T∗:H2→H1T^* : H_2 \to H_1 is the unique operator satisfying ⟨T(x),y⟩H2=⟨x,T∗(y)⟩H1\langle T(x), y \rangle_{H_2} = \langle x, T^*(y) \rangle_{H_1} for all x∈H1x \in H_1 and y∈H2y \in H_2.

The existence of T∗T^* follows from the Riesz Representation Theorem.

ExampleComputing Adjoints
  1. Matrix Operators: For finite-dimensional spaces, T∗T^* corresponds to the conjugate transpose of the matrix representing TT

  2. Multiplication Operator: If (Mϕf)(x)=ϕ(x)f(x)(M_\phi f)(x) = \phi(x) f(x) on L2L^2, then (Mϕ∗f)(x)=ϕ(x)‾f(x)(M_\phi^* f)(x) = \overline{\phi(x)} f(x)

  3. Integral Operator: If (Kf)(x)=∫abk(x,y)f(y) dy(Kf)(x) = \int_a^b k(x,y) f(y) \, dy on L2[a,b]L^2[a,b], then (K∗f)(x)=∫abk(y,x)‾f(y) dy(K^* f)(x) = \int_a^b \overline{k(y,x)} f(y) \, dy

  4. Shift Operator: On ℓ2\ell^2, if (Sx)n=xn+1(Sx)_n = x_{n+1}, then (S∗x)n=xn−1(S^*x)_n = x_{n-1} (with x0=0x_0 = 0)

DefinitionSpecial Classes of Operators

Let T∈B(H)T \in \mathcal{B}(H) where HH is a Hilbert space:

  1. Self-adjoint: T=T∗T = T^*
  2. Normal: TT∗=T∗TTT^* = T^*T
  3. Unitary: T∗T=TT∗=IT^*T = TT^* = I (and TT is surjective)
  4. Positive: ⟨Tx,x⟩≥0\langle Tx, x \rangle \geq 0 for all xx
  5. Projection: T2=T=T∗T^2 = T = T^*
TheoremProperties of the Adjoint

Let S,T∈B(H1,H2)S, T \in \mathcal{B}(H_1, H_2) and R∈B(H2,H3)R \in \mathcal{B}(H_2, H_3). Then:

  1. (S+T)∗=S∗+T∗(S + T)^* = S^* + T^*
  2. (αT)∗=α‾T∗(\alpha T)^* = \overline{\alpha} T^*
  3. (RT)∗=T∗R∗(RT)^* = T^*R^*
  4. (T∗)∗=T(T^*)^* = T
  5. ∥T∗T∥=∥T∥2\|T^*T\| = \|T\|^2
  6. ker⁡(T∗)=(Range(T))⊥\ker(T^*) = (\text{Range}(T))^\perp
Remark

Self-adjoint operators play a role analogous to symmetric matrices in finite dimensions. They have real spectra and admit spectral decompositions, making them essential in quantum mechanics where observables are represented by self-adjoint operators.

The adjoint operation provides a rich algebraic structure on B(H)\mathcal{B}(H), turning it into a C∗C^*-algebra. This structure is fundamental to operator theory and has applications ranging from quantum mechanics to signal processing.