TheoremBase

The embedding is bounded by condition (a); the Riesz map comes from the Riesz representation theorem on V applied to the functional <z,.>_H, and its properties follow from uniqueness there; D(A) = J(H) by comparing the defining identities; separability passes to H because a set dense in V is dense in H.

Proof

We use Hilbert Triple: a Densely and Continuously Embedded Hilbert Space and Its Form Operator, the identities of Elementary Identities in a Real Inner Product Space, The Cauchy–Schwarz Inequality in a Real Inner Product Space, Elementary Properties of Bounded Linear Maps and Functionals on Real Inner Product Spaces and The Riesz Representation Theorem for a Real Hilbert Space. Condition (a) of the triple is ∣x∣H≤∣x∣V|x|_{H}\le|x|_{V} for x∈Vx\in V, and condition (b) that VV is dense in HH.

Claim 1. ι\iota is linear since the operations of VV are those of HH (claim 1 of A Linear Subspace is a Vector Space and Inherits an Inner Product), and ∣ιx∣H=∣x∣H≤1⋅∣x∣V|\iota x|_{H}=|x|_{H}\le 1\cdot|x|_{V}, so ι\iota is bounded with operator norm at most 11 by Elementary Properties of Bounded Linear Maps and Functionals on Real Inner Product Spaces (i). If (xm)(x_{m}) in VV converges to xx in VV, then for every ε>0\varepsilon>0 there is NN with ∣xm−x∣V<ε|x_{m}-x|_{V}<\varepsilon for m≥Nm\ge N, hence ∣xm−x∣H≤∣xm−x∣V<ε|x_{m}-x|_{H}\le|x_{m}-x|_{V}<\varepsilon; so (xm)(x_{m}) converges to xx in HH (Convergent Sequence in a Metric Space). The Cauchy statement is identical with ∣xm−xn∣|x_{m}-x_{n}| in place of ∣xm−x∣|x_{m}-x| (Cauchy Sequence in a Metric Space).

Claim 2. Let z∈Hz\in H and ℓz(y)=⟨z,y⟩H\ell_{z}(y)=\langle z,y\rangle_{H} for y∈Vy\in V. It is linear by Elementary Identities in a Real Inner Product Space (i), and ∣ℓz(y)∣≤∣z∣H∣y∣H≤∣z∣H∣y∣V|\ell_{z}(y)|\le|z|_{H}|y|_{H}\le|z|_{H}|y|_{V} by Cauchy-Schwarz, condition (a) and claim 5 of Elementary Arithmetic in an Ordered Field; so ℓz\ell_{z} is a bounded linear functional on VV and ∥ℓz∥≤∣z∣H\lVert\ell_{z}\rVert\le|z|_{H} by Elementary Properties of Bounded Linear Maps and Functionals on Real Inner Product Spaces (i) (via Elementary Properties of Bounded Linear Maps and Functionals on Real Inner Product Spaces (viii)). Since VV is a real Hilbert space, The Riesz Representation Theorem for a Real Hilbert Space (i) and The Riesz Representation Theorem for a Real Hilbert Space (ii) give a unique Jz∈VJz\in V with ⟨y,Jz⟩V=ℓz(y)\langle y,Jz\rangle_{V}=\ell_{z}(y), that is (by symmetry) ⟨Jz,y⟩V=⟨z,y⟩H\langle Jz,y\rangle_{V}=\langle z,y\rangle_{H}, for every y∈Vy\in V; and ∣Jz∣V=∥ℓz∥≤∣z∣H|Jz|_{V}=\lVert\ell_{z}\rVert\le|z|_{H} by The Riesz Representation Theorem for a Real Hilbert Space (iii).

Linearity. For z,z′∈Hz,z'\in H, λ∈R\lambda\in\mathbb{R} and y∈Vy\in V: ⟨Jz+Jz′,y⟩V=⟨z,y⟩H+⟨z′,y⟩H=⟨z+z′,y⟩H=⟨J(z+z′),y⟩V\langle Jz+Jz',y\rangle_{V}=\langle z,y\rangle_{H}+\langle z',y\rangle_{H}=\langle z+z',y\rangle_{H}=\langle J(z+z'),y\rangle_{V}, so J(z+z′)=Jz+Jz′J(z+z')=Jz+Jz' by the uniqueness clause; likewise J(λz)=λJzJ(\lambda z)=\lambda Jz.

Identities. ⟨Jz,Jz′⟩V=⟨z,Jz′⟩H\langle Jz,Jz'\rangle_{V}=\langle z,Jz'\rangle_{H} by the defining identity of JzJz with y=Jz′∈Vy=Jz'\in V; and ⟨Jz,z′⟩H=⟨z′,Jz⟩H=⟨Jz′,Jz⟩V=⟨Jz,Jz′⟩V\langle Jz,z'\rangle_{H}=\langle z',Jz\rangle_{H}=\langle Jz',Jz\rangle_{V}=\langle Jz,Jz'\rangle_{V} by symmetry and the defining identity of Jz′Jz' with y=Jzy=Jz.

Injectivity. If Jz=0HJz=0_{H} (the zero vector of VV by claim 1 of A Linear Subspace is a Vector Space and Inherits an Inner Product), then ⟨z,y⟩H=⟨Jz,y⟩V=⟨0H,y⟩V=0\langle z,y\rangle_{H}=\langle Jz,y\rangle_{V}=\langle 0_{H},y\rangle_{V}=0 for every y∈Vy\in V (Elementary Identities in a Real Inner Product Space (ii)), so z∈V⊥z\in V^{\perp}, which is {0H}\{0_{H}\} by Orthogonal Projection onto a Closed Linear Subspace of a Real Hilbert Space (vi) and condition (b); hence z=0Hz=0_{H}, and JJ being linear, it is injective.

Weak convergence. Let (xm)(x_{m}) in VV converge weakly to x∈Vx\in V in VV, and let z∈Hz\in H. Then ⟨xm,z⟩H=⟨z,xm⟩H=⟨Jz,xm⟩V=⟨xm,Jz⟩V\langle x_{m},z\rangle_{H}=\langle z,x_{m}\rangle_{H}=\langle Jz,x_{m}\rangle_{V}=\langle x_{m},Jz\rangle_{V}, which converges to ⟨x,Jz⟩V=⟨x,z⟩H\langle x,Jz\rangle_{V}=\langle x,z\rangle_{H} by Weak Convergence of a Sequence in a Real Inner Product Space applied with the vector Jz∈VJz\in V. As zz was arbitrary, (xm)(x_{m}) converges weakly to xx in HH.

Claim 3. If x∈D(A)x\in D(A) and z=Axz=Ax, then ⟨x,y⟩V=⟨z,y⟩H=⟨Jz,y⟩V\langle x,y\rangle_{V}=\langle z,y\rangle_{H}=\langle Jz,y\rangle_{V} for every y∈Vy\in V, so x=Jz=J(Ax)x=Jz=J(Ax) by the uniqueness in claim 2 (or by Elementary Identities in a Real Inner Product Space (iii) with y=x−Jzy=x-Jz). Conversely, for z∈Hz\in H the vector x=Jzx=Jz satisfies ⟨x,y⟩V=⟨z,y⟩H\langle x,y\rangle_{V}=\langle z,y\rangle_{H} for all y∈Vy\in V, so x∈D(A)x\in D(A) with Ax=zAx=z by Hilbert Triple: a Densely and Continuously Embedded Hilbert Space and Its Form Operator (iv); that is, A(Jz)=zA(Jz)=z. Hence D(A)=J(H)D(A)=J(H). Since JJ is linear, J(H)J(H) contains 0H=J0H0_{H}=J0_{H} and is closed under sums and scalar multiples, so D(A)D(A) is a linear subspace of VV; and A(Jz+Jz′)=A(J(z+z′))=z+z′=A(Jz)+A(Jz′)A(Jz+Jz')=A(J(z+z'))=z+z'=A(Jz)+A(Jz'), A(λJz)=A(J(λz))=λz=λA(Jz)A(\lambda Jz)=A(J(\lambda z))=\lambda z=\lambda A(Jz), so AA is linear.

Claim 4. The first identity is Hilbert Triple: a Densely and Continuously Embedded Hilbert Space and Its Form Operator (iv). For x,y∈D(A)x,y\in D(A), by it and symmetry, ⟨Ax,y⟩H=⟨x,y⟩V=⟨y,x⟩V=⟨Ay,x⟩H=⟨x,Ay⟩H\langle Ax,y\rangle_{H}=\langle x,y\rangle_{V}=\langle y,x\rangle_{V}=\langle Ay,x\rangle_{H}=\langle x,Ay\rangle_{H}. With y=xy=x, ⟨Ax,x⟩H=⟨x,x⟩V=∣x∣V2≥∣x∣H2\langle Ax,x\rangle_{H}=\langle x,x\rangle_{V}=|x|_{V}^{2}\ge|x|_{H}^{2} by condition (a) and claim 2 of Monotonicity of Squaring on the Nonnegative Elements of an Ordered Field.

Claim 5. Let D⊆VD\subseteq V be countable and dense in (V,dV)(V,d_{V}), and let Dˉ\bar{D} be its closure in (H,dH)(H,d_{H}), a closed subset of HH containing DD by claims 1 and 2 of The Closure is the Smallest Closed Superset. If x∈Vx\in V, then by Sequential Characterization of the Closure in a Metric Space (in VV) there is a sequence in DD converging to xx in VV, hence in HH by claim 1, so x∈Dˉx\in\bar{D} by the same lemma in HH. Thus V⊆DˉV\subseteq\bar{D}, and by claim 3 of The Closure is the Smallest Closed Superset the closure of VV in HH, which is HH by condition (b), is contained in Dˉ\bar{D}. So Dˉ=H\bar{D}=H: DD is a countable dense subset of (H,dH)(H,d_{H}), which is therefore separable.

Citations

Loading…

Dependencies

Uses0

Loading…

Comments

Log in to comment.

Loading…