TheoremBase

A dyadic simple minorant bounded by N carries all but ε/2 of the integral by monotone convergence; on a set of measure below ε/(2N) the minorant contributes less than ε/2. The integrable case follows by applying this to |f|.

Proof

Proof of (i). Let ε>0\varepsilon>0. Since ∫Xf dμ<∞\int_{X}f\,d\mu<\infty, this integral is a real number, and so is every integral of an element of M+\mathcal{M}^{+} below, each being at most ∫Xf dμ\int_{X}f\,d\mu (Small Facts about [0,∞]: Zero, Finiteness and Sums of Zeros (ii) and Small Facts about [0,∞]: Zero, Finiteness and Sums of Zeros (iii), with Basic Properties of the Integral of Nonnegative Measurable Maps (iii), and with Integrals over Measurable Subsets (viii) and Integrals over Measurable Subsets (v) for ∫Af dμ≤∫Xf dμ\int_{A}f\,d\mu\le\int_{X}f\,d\mu when A∈FA\in\mathcal{F}).

Step 1: a bounded minorant carrying almost all of the integral. For m∈Nm\in\mathbb{N} let sm=sm(f)s_{m}=s_{m}(f) be the dyadic approximation of ff of order mm. By Approximation of Measurable Functions by Dyadic Simple Functions (i), sm∈S+⊆M+s_{m}\in\mathcal{S}^{+}\subseteq\mathcal{M}^{+} (Basic Properties of the Integral of Nonnegative Measurable Maps (i)), sm(x)≤sm+1(x)≤f(x)s_{m}(x)\le s_{m+1}(x)\le f(x) and f(x)=sup⁡msm(x)f(x)=\sup_{m}s_{m}(x) for every x∈Xx\in X; moreover sm(x)≤ms_{m}(x)\le m for every x∈Xx\in X, by Dyadic Approximations of a Map into [0,∞] (i). By the monotone convergence theorem,

∫Xf dμ=sup⁡m∫Xsm dμ.\int_{X}f\,d\mu=\sup_{m}\int_{X}s_{m}\,d\mu .

Since ∫Xf dμ−ε/2\int_{X}f\,d\mu-\varepsilon/2 is less than this supremum, there is N∈NN\in\mathbb{N} with

∫Xf dμ−∫XsN dμ<ε2\int_{X}f\,d\mu-\int_{X}s_{N}\,d\mu<\frac{\varepsilon}{2}

(Order and Suprema in the Extended Real Line (x)). First fix such an NN, which depends only on ff and ε\varepsilon; then put

δ=ε2N>0\delta=\frac{\varepsilon}{2N}>0

(N≥1N\ge1 in R\mathbb{R} by Natural Numbers and Factorials Read in the Reals: the Order Carries Over, One Is Least, and Factorials Are Positive, Recursive and Nondecreasing (ii), so δ\delta is defined and positive, and N≥0N\ge0).

Step 2: the estimate. Let A∈FA\in\mathcal{F} with μ(A)<δ\mu(A)<\delta, so that μ(A)<∞\mu(A)<\infty and μ(A)\mu(A) is a real number (Small Facts about [0,∞]: Zero, Finiteness and Sums of Zeros (ii)), and let B=X∖A∈FB=X\setminus A\in\mathcal{F}. As AA and BB are disjoint with union XX, additivity over disjoint sets gives, for h=fh=f and for h=sNh=s_{N},

∫Xh dμ=∫Ah dμ+∫Bh dμ\int_{X}h\,d\mu=\int_{A}h\,d\mu+\int_{B}h\,d\mu

(Integrals over Measurable Subsets (vii) and Integrals over Measurable Subsets (v)). Writing integrals over a set as integrals of products with its indicator (Integrals over Measurable Subsets (i)), the pointwise bounds sN1B≤f1Bs_{N}\mathbf{1}_{B}\le f\mathbf{1}_{B} and sN1A≤N1As_{N}\mathbf{1}_{A}\le N\mathbf{1}_{A} give

∫BsN dμ≤∫Bf dμ,∫AsN dμ≤∫XN1A dμ=Nμ(A)\int_{B}s_{N}\,d\mu\le\int_{B}f\,d\mu,\qquad\int_{A}s_{N}\,d\mu\le\int_{X}N\mathbf{1}_{A}\,d\mu=N\mu(A)

(Basic Properties of the Integral of Nonnegative Measurable Maps (iii), Basic Properties of the Integral of Nonnegative Measurable Maps (iv) and Basic Properties of the Integral of Nonnegative Measurable Maps (ii)). Hence

∫Af dμ=∫Xf dμ−∫Bf dμ≤∫Xf dμ−∫BsN dμ=(∫Xf dμ−∫XsN dμ)+∫AsN dμ<ε2+Nδ=ε.\int_{A}f\,d\mu=\int_{X}f\,d\mu-\int_{B}f\,d\mu\le\int_{X}f\,d\mu-\int_{B}s_{N}\,d\mu=\Bigl(\int_{X}f\,d\mu-\int_{X}s_{N}\,d\mu\Bigr)+\int_{A}s_{N}\,d\mu<\frac{\varepsilon}{2}+N\delta=\varepsilon .

Proof of (ii). Here ∣f∣∈L1(μ)|f|\in\mathcal{L}^{1}(\mu) by Linearity and Monotonicity of the Lebesgue Integral, and the Integrability Criterion (x), and ff and ∣f∣|f| are integrable on every A∈FA\in\mathcal{F} by Integrals over Measurable Subsets (vi), so the integrals in (ii) are defined. Let ε>0\varepsilon>0. By Linearity and Monotonicity of the Lebesgue Integral, and the Integrability Criterion (iv), ∣f∣∈M+|f|\in\mathcal{M}^{+} and ∫X∣f∣ dμ<∞\int_{X}|f|\,d\mu<\infty, so (i), applied to ∣f∣|f| and ε\varepsilon, gives δ>0\delta>0 such that ∫A∣f∣ dμ<ε\int_{A}|f|\,d\mu<\varepsilon for every A∈FA\in\mathcal{F} with μ(A)<δ\mu(A)<\delta; this integral of ∣f∣∈M+|f|\in\mathcal{M}^{+} over AA equals that of ∣f∣|f| as a function integrable on AA (Integrals over Measurable Subsets (iii)). For such AA, let (A,FA,μA)(A,\mathcal{F}_{A},\mu_{A}) be the restriction of (X,F,μ)(X,\mathcal{F},\mu) to AA, a measure space (Integrals of Functions over a Measurable Subset (i)). Since ∣f∣A∣=∣f∣∣A\lvert f|_{A}\rvert=\lvert f\rvert|_{A}, the definition of integrals over AA (Integrals of Functions over a Measurable Subset (iv)) and the absolute-value bound in (A,FA,μA)(A,\mathcal{F}_{A},\mu_{A}) (Linearity and Monotonicity of the Lebesgue Integral, and the Integrability Criterion (x)) give

∣∫Af dμ∣=∣∫f∣A dμA∣≤∫∣f∣A∣ dμA=∫A∣f∣ dμ<ε.\Bigl|\int_{A}f\,d\mu\Bigr|=\Bigl|\int f|_{A}\,d\mu_{A}\Bigr|\le\int\lvert f|_{A}\rvert\,d\mu_{A}=\int_{A}|f|\,d\mu<\varepsilon .

Citations

Loading…

Dependencies

Uses0

Loading…

Comments

Log in to comment.

Loading…