TheoremBase

Proof

Throughout, λ=λ[0,T]\lambda=\lambda_{[0,T]} and B=B[0,T]\mathcal{B}=\mathcal{B}_{[0,T]} are as in the definition of the Lebesgue space, with λ([0,T])=T\lambda([0,T])=T by the interval toolkit, and the norm of a class does not depend on the representative used to compute it, by The Lebesgue Space of Square-Integrable Vector-Valued Functions is a Real Inner Product Space. Fix a0∈Aa_{0}\in\mathcal{A}, which is possible since A\mathcal{A} is nonempty.

A normalized representative. We first record: every ξ∈UA\xi\in\mathcal{U}_{\mathcal{A}} has a representative taking all its values in A\mathcal{A}. Indeed, let u∈L2([0,T];Rm)u\in\mathcal{L}^{2}([0,T];\mathbb{R}^{m}) represent ξ\xi and let N∈BN\in\mathcal{B} with λ(N)=0\lambda(N)=0 be such that u(t)∈Au(t)\in\mathcal{A} for t∈[0,T]∖Nt\in[0,T]\setminus N. Define u~(t)=u(t)\tilde u(t)=u(t) for t∉Nt\notin N and u~(t)=a0\tilde u(t)=a_{0} for t∈Nt\in N; each component u~i\tilde u^{i} is measurable: for a set BB in the Borel σ\sigma-algebra of R\mathbb{R},

(u~i)−1(B)=((ui)−1(B)∖N)∪NB,(\tilde u^{i})^{-1}(B)=\bigl((u^{i})^{-1}(B)\setminus N\bigr)\cup N_{B},

where NB=NN_{B}=N if a0i∈Ba_{0}^{i}\in B and NB=∅N_{B}=\emptyset otherwise; here (ui)−1(B)∈B(u^{i})^{-1}(B)\in\mathcal{B} because uiu^{i} is measurable, N∈BN\in\mathcal{B}, and B\mathcal{B} is closed under differences and finite unions by the definition of a σ\sigma-algebra. Moreover u~=u\tilde u=u off the null set NN, so u~∼u\tilde u\sim u, hence u~\tilde u is square-integrable and [u~]=ξ[\tilde u]=\xi; and u~(t)∈A\tilde u(t)\in\mathcal{A} for every tt.

Claim 1. Let a∈Aa\in\mathcal{A} and let uu be the constant map with value aa. Each component uiu^{i} is constant, hence a simple function on ([0,T],B)([0,T],\mathcal{B}) and in particular measurable, and ∣u(t)∣2=∣a∣2|u(t)|^{2}=|a|^{2} for every tt, so ∫[0,T]∣u∣2 dλ=∣a∣2T<∞\int_{[0,T]}|u|^{2}\,d\lambda=|a|^{2}T<\infty. Thus u∈L2([0,T];Rm)u\in\mathcal{L}^{2}([0,T];\mathbb{R}^{m}), and u(t)∈Au(t)\in\mathcal{A} for every tt, so [u]∈UA[u]\in\mathcal{U}_{\mathcal{A}} (take N=∅N=\emptyset).

Claim 2. Let ξ∈UA\xi\in\mathcal{U}_{\mathcal{A}} and let u~\tilde u be a representative with all values in A\mathcal{A}. Then ∣u~(t)∣2≤R2|\tilde u(t)|^{2}\le R^{2} for every tt, so by monotonicity of the integral (claim 1 of Linearity and Monotonicity of the Lebesgue Integral)

∥ξ∥L22=∫[0,T]∣u~∣2 dλ≤R2T.\lVert\xi\rVert^{2}_{L^{2}}=\int_{[0,T]}|\tilde u|^{2}\,d\lambda\le R^{2}T .

Both ∥ξ∥L2\lVert\xi\rVert_{L^{2}} and RT1/2RT^{1/2} are nonnegative and their squares satisfy ∥ξ∥L22≤(RT1/2)2\lVert\xi\rVert_{L^{2}}^{2}\le(RT^{1/2})^{2}, so ∥ξ∥L2≤RT1/2\lVert\xi\rVert_{L^{2}}\le RT^{1/2}: otherwise RT1/2<∥ξ∥L2RT^{1/2}<\lVert\xi\rVert_{L^{2}}, and multiplying this inequality by the two nonnegative numbers in turn would give (RT1/2)2<∥ξ∥L22(RT^{1/2})^{2}<\lVert\xi\rVert^{2}_{L^{2}}.

Claim 3. Let ξ,ξ′∈UA\xi,\xi'\in\mathcal{U}_{\mathcal{A}} and let ss be real with 0≤s≤10\le s\le1. Choose representatives u~,u~′\tilde u,\tilde u' with all values in A\mathcal{A}. By the operations of The Lebesgue Space of Square-Integrable Vector-Valued Functions on a Compact Interval, su~+(1−s)u~′s\tilde u+(1-s)\tilde u' is a representative of sξ+(1−s)ξ′s\xi+(1-s)\xi', and it is square-integrable because L2([0,T];Rm)L^{2}([0,T];\mathbb{R}^{m}) is a real vector space with these operations by The Lebesgue Space of Square-Integrable Vector-Valued Functions is a Real Inner Product Space. By the hypothesis of this claim, su~(t)+(1−s)u~′(t)∈As\tilde u(t)+(1-s)\tilde u'(t)\in\mathcal{A} for every tt. Hence sξ+(1−s)ξ′∈UAs\xi+(1-s)\xi'\in\mathcal{U}_{\mathcal{A}}, which is convexity.

Claim 4. The metric open subsets of HH determined by dL2d_{L^{2}} form a topology by Metric Open Sets Form a Topology, and likewise for Rm\mathbb{R}^{m} with the Euclidean distance dEd_{E}. Let ξ\xi belong to the closure of UA\mathcal{U}_{\mathcal{A}} in (H,dL2)(H,d_{L^{2}}).

For n∈Nn\in\mathbb{N} let

Bn={v∈L2([0,T];Rm)  :  v(t)∈A for every t∈[0,T] and dL2(ξ,[v])<ι(n)−1},B_{n}=\bigl\{v\in\mathcal{L}^{2}([0,T];\mathbb{R}^{m})\;:\;v(t)\in\mathcal{A}\text{ for every }t\in[0,T]\text{ and }d_{L^{2}}(\xi,[v])<\iota(n)^{-1}\bigr\},

where ι\iota is the canonical map of R\mathbb{R}, so that ι(n)−1>0\iota(n)^{-1}>0 by claim 3 of Properties of the Canonical Map from the Natural Numbers to an Ordered Field. By Characterization of the Closure in a Metric Space by Open Balls there is an element of UA\mathcal{U}_{\mathcal{A}} within distance ι(n)−1\iota(n)^{-1} of ξ\xi, and by the normalized-representative paragraph it has a representative with all values in A\mathcal{A}; hence Bn≠∅B_{n}\ne\emptyset. The sets BnB_{n} are subsets of L2([0,T];Rm)\mathcal{L}^{2}([0,T];\mathbb{R}^{m}), so Axiom of Countable Choice provides a sequence (un)n∈N(u_{n})_{n\in\mathbb{N}} with un∈Bnu_{n}\in B_{n} for every nn.

Then 0≤∥[un]−ξ∥L2<ι(n)−10\le\lVert[u_{n}]-\xi\rVert_{L^{2}}<\iota(n)^{-1}, and the real sequence (ι(n)−1)n(\iota(n)^{-1})_{n} has limit 00 by claim 3 of The Archimedean Property of the Real Numbers; comparing with the constant sequence 00 and using Order Properties of Limits of Real Sequences, the real sequence (∥[un]−ξ∥L2)n\bigl(\lVert[u_{n}]-\xi\rVert_{L^{2}}\bigr)_{n} has limit 00.

Let uu be a representative of ξ\xi. By claim 2 of Completeness of the Lebesgue Space of Square-Integrable Vector-Valued Functions there are natural numbers n1<n2<…n_{1}<n_{2}<\dots and a set N∈BN\in\mathcal{B} with λ(N)=0\lambda(N)=0 such that (unj(t))j\bigl(u_{n_{j}}(t)\bigr)_{j} converges to u(t)u(t) in (Rm,dE)(\mathbb{R}^{m},d_{E}) for every t∈[0,T]∖Nt\in[0,T]\setminus N. Fix such a tt and let ε>0\varepsilon>0 be real; there is jj with dE(u(t),unj(t))<εd_{E}(u(t),u_{n_{j}}(t))<\varepsilon, and unj(t)∈Au_{n_{j}}(t)\in\mathcal{A}. By Characterization of the Closure in a Metric Space by Open Balls, u(t)u(t) lies in the closure of A\mathcal{A} in (Rm,dE)(\mathbb{R}^{m},d_{E}); since A\mathcal{A} is closed it equals its closure, so u(t)∈Au(t)\in\mathcal{A}.

Thus u(t)∈Au(t)\in\mathcal{A} for every t∈[0,T]∖Nt\in[0,T]\setminus N with λ(N)=0\lambda(N)=0, so ξ=[u]∈UA\xi=[u]\in\mathcal{U}_{\mathcal{A}}. Therefore the closure of UA\mathcal{U}_{\mathcal{A}} is contained in UA\mathcal{U}_{\mathcal{A}}; as it also contains UA\mathcal{U}_{\mathcal{A}} and is closed, UA\mathcal{U}_{\mathcal{A}} equals its closure and is closed.

Citations

Loading…

Dependencies

Uses0

Loading…

Comments

Log in to comment.

Loading…