TheoremBase

Proof sketch: the compact case is the Riemann–Lebesgue agreement on the restricted space followed by zero extension; then the truncations of h increase to h, so monotone convergence gives the supremum, and a bounded nondecreasing real sequence converges to its supremum.

Proof

Throughout, closed intervals are Borel sets (background of the settings), h∣[a,b]h|_{[a,b]} is continuous on [a,b][a,b] when hh is continuous on R\mathbb{R} (The Real Line: Standing Notation and Background for Calculus on Intervals (vi)), and the functions h1[−n,n]h\mathbf{1}_{[-n,n]} form a sequence of functions R→R\mathbb{R}\to\mathbb{R} (Intervals of the Real Line as Defined Set Symbols, Typed Terms and Values of Maps (iii)).

Step 1 (proof of (i)). Let φ=g∣[a,b]\varphi=g|_{[a,b]}. By Restricted Lebesgue Measure and Integral Toolkit on a Compact Interval (iii), φ\varphi is measurable for B[a,b]\mathcal{B}_{[a,b]}, φ∈L1(λ[a,b])\varphi\in\mathcal{L}^{1}(\lambda_{[a,b]}) and ∫φ dλ[a,b]=∫abφ\int\varphi\,d\lambda_{[a,b]}=\int_{a}^{b}\varphi; here ([a,b],B[a,b],λ[a,b])([a,b],\mathcal{B}_{[a,b]},\lambda_{[a,b]}) is the restriction of (R,B(R),λ)(\mathbb{R},\mathcal{B}(\mathbb{R}),\lambda) to [a,b][a,b] of Integrals of Functions over a Measurable Subset (i). The function g1[a,b]g\mathbf{1}_{[a,b]} equals φ\varphi on [a,b][a,b] and 00 off [a,b][a,b], so by Integrals over Measurable Subsets (iv) it is measurable, lies in L1(λ)\mathcal{L}^{1}(\lambda), and

∫Rg1[a,b] dλ=∫φ dλ[a,b]=∫abg∣[a,b].\int_{\mathbb{R}}g\mathbf{1}_{[a,b]}\,d\lambda=\int\varphi\,d\lambda_{[a,b]}=\int_{a}^{b}g|_{[a,b]} .

Step 2 (proof of (ii)). Let fn=h1[−n,n]f_{n}=h\mathbf{1}_{[-n,n]} for n∈Nn\in\mathbb{N}. Since 0<n0<n in R\mathbb{R} (Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares (vi)), −n<n-n<n, so Step 1 applies with g=hg=h, a=−na=-n and b=nb=n: fnf_{n} is measurable and ∫Rfn dλ=In\int_{\mathbb{R}}f_{n}\,d\lambda=I_{n}. As fnf_{n} is nonnegative, Linearity and Monotonicity of the Lebesgue Integral, and the Integrability Criterion (vi) puts it in M+(R,B(R))\mathcal{M}^{+}(\mathbb{R},\mathcal{B}(\mathbb{R})) with

∫Rfn dλ=In∈[0,∞)\int_{\mathbb{R}}f_{n}\,d\lambda=I_{n}\in[0,\infty)

for its integral as an element of M+\mathcal{M}^{+}; so the sequence (In)(I_{n}) whose supremum (ii) takes is the sequence n↦∫Rfn dλn\mapsto\int_{\mathbb{R}}f_{n}\,d\lambda of these integrals. Since h≥0h\ge0 and [−n,n]⊆[−n−1,n+1][-n,n]\subseteq[-n-1,n+1], fn(x)≤fn+1(x)≤h(x)f_{n}(x)\le f_{n+1}(x)\le h(x) for every x∈Rx\in\mathbb{R}. Given x∈Rx\in\mathbb{R}, the Archimedean property (Infima, the Archimedean Property, Density of the Rationals and Rational Approximation from Below in an Ordered Field Whose Nonempty Sets Bounded Above Have Suprema (ii)) gives N1,N2∈NN_{1},N_{2}\in\mathbb{N} with x<N1x<N_{1} and −x<N2-x<N_{2}; for N=N1+N2N=N_{1}+N_{2} we have −N<x<N-N<x<N, as N1,N2>0N_{1},N_{2}>0, so x∈[−N,N]x\in[-N,N] and fN(x)=h(x)f_{N}(x)=h(x). Hence h(x)=sup⁡nfn(x)h(x)=\sup_{n}f_{n}(x) for every x∈Rx\in\mathbb{R}. By the monotone convergence theorem, h∈M+(R,B(R))h\in\mathcal{M}^{+}(\mathbb{R},\mathcal{B}(\mathbb{R})) (Monotone Convergence Theorem (i)) and, by Monotone Convergence Theorem (ii),

∫Rh dλ=sup⁡n∈N∫Rfn dλ=sup⁡n∈NIn.\int_{\mathbb{R}}h\,d\lambda=\sup_{n\in\mathbb{N}}\int_{\mathbb{R}}f_{n}\,d\lambda=\sup_{n\in\mathbb{N}}I_{n} .

Step 3 (proof of (iii)). The function hh is real-valued, nonnegative and lies in M+\mathcal{M}^{+} by Step 2, so it is measurable as a map into R\mathbb{R} (Measurable Maps: Generators, Composition, Continuous Maps, Restrictions, and Criteria for Real- and [0,∞]-Valued Maps (vii)). By Linearity and Monotonicity of the Lebesgue Integral, and the Integrability Criterion (vi), h∈L1(λ)h\in\mathcal{L}^{1}(\lambda) if and only if ∫Rh dλ<∞\int_{\mathbb{R}}h\,d\lambda<\infty for the integral of h∈M+h\in\mathcal{M}^{+}, and then the two integrals of hh agree; by Step 2 this is sup⁡nIn<∞\sup_{n}I_{n}<\infty.

Step 4 (proof of (iv)). Let h∈L1(λ)h\in\mathcal{L}^{1}(\lambda). By Step 3, ∫Rh dλ\int_{\mathbb{R}}h\,d\lambda, the integral of h∈M+h\in\mathcal{M}^{+}, is a real number and equals the integral of h∈L1(λ)h\in\mathcal{L}^{1}(\lambda). The sequence (In)(I_{n}) is nondecreasing, since In=∫fn dλ≤∫fn+1 dλ=In+1I_{n}=\int f_{n}\,d\lambda\le\int f_{n+1}\,d\lambda=I_{n+1} by Basic Properties of the Integral of Nonnegative Measurable Maps (iii), and bounded above by ∫Rh dλ\int_{\mathbb{R}}h\,d\lambda, since fn≤hf_{n}\le h gives In≤∫Rh dλI_{n}\le\int_{\mathbb{R}}h\,d\lambda by the same monotonicity. By Completeness of the Real Numbers for Sequences: Monotone Convergence, the Bolzano-Weierstrass Theorem and Cauchy Sequences (iii), InI_{n} converges to the supremum of {In:n∈N}\{I_{n}:n\in\mathbb{N}\} in R\mathbb{R}, which is sup⁡nIn\sup_{n}I_{n} by Order and Suprema in the Extended Real Line (iv), and so ∫Rh dλ\int_{\mathbb{R}}h\,d\lambda by Step 2.

Citations

Loading…

Dependencies

Uses0

Loading…

Comments

Log in to comment.

Loading…