TheoremBase

Each image measure in question is a Borel measure giving every nonempty open box its volume, hence is Lebesgue measure by uniqueness; the integral formulas follow by moving integrals through concatenation and applying Tonelli and Fubini to f∘ι.

Proof

Throughout, products of nonnegative real numbers in [0,∞][0,\infty] are their real products (Addition and Multiplication on [0,∞] (ii)). A measure on (Rp,B(Rp))(\mathbb{R}^{p},\mathcal{B}(\mathbb{R}^{p})), p∈Np\in\mathbb{N}, that gives every open box R(a,b)R(a,b) with ai<bia_{i}<b_{i} for all i∈[p]i\in[p] the value ∏i=1p(bi−ai)\prod_{i=1}^{p}(b_{i}-a_{i}) is λp\lambda_{p}, by the uniqueness in Existence and Uniqueness of Lebesgue Measure on ℝⁿ (iii) and Lebesgue Measure on Rn\mathbb{R}^n (i).

Dimension one. By Borel Sets of ℝ^{m+n} and the Product σ-Algebra of ℝ^m and ℝ^n (iii), τ\tau is a B(R)/B(R1)\mathcal{B}(\mathbb{R})/\mathcal{B}(\mathbb{R}^{1})-measurable bijection, and τ[B]∈B(R1)\tau[B]\in\mathcal{B}(\mathbb{R}^{1}) for every B∈B(R)B\in\mathcal{B}(\mathbb{R}). The image measure λ∘τ−1\lambda\circ\tau^{-1} is a measure on (R1,B(R1))(\mathbb{R}^{1},\mathcal{B}(\mathbb{R}^{1})) (Image Measures and Measures with a Density: Measures, Integrals and Change of Variables (i), Lebesgue Measure on the Real Line: Existence, Lengths of Intervals and Uniqueness (iv)). For a,b∈R1a,b\in\mathbb{R}^{1} with a1<b1a_{1}<b_{1} we have τ−1(R(a,b))=(a1,b1)\tau^{-1}(R(a,b))=(a_{1},b_{1}), since τ(t)1=t\tau(t)_{1}=t, so

(λ∘τ−1)(R(a,b))=λ((a1,b1))=b1−a1=∏i=11(bi−ai)(\lambda\circ\tau^{-1})\bigl(R(a,b)\bigr)=\lambda\bigl((a_{1},b_{1})\bigr)=b_{1}-a_{1}=\prod_{i=1}^{1}(b_{i}-a_{i})

(Lebesgue Measure on the Real Line: Existence, Lengths of Intervals and Uniqueness (i), Iterated Operations: Recursion, Splitting, Reordering, Termwise Combination and Homomorphisms (i)). Hence λ∘τ−1=λ1\lambda\circ\tau^{-1}=\lambda_{1}, and for every B∈B(R)B\in\mathcal{B}(\mathbb{R}), since τ−1(τ[B])=B\tau^{-1}(\tau[B])=B for the bijection τ\tau (Images and Preimages under a Bijection and Its Inverse (ii)),

λ1(τ[B])=λ(τ−1(τ[B]))=λ(B).\lambda_{1}(\tau[B])=\lambda\bigl(\tau^{-1}(\tau[B])\bigr)=\lambda(B).

This proves (i).

Concatenation. The measures λm\lambda_{m} and λn\lambda_{n} are σ\sigma-finite (Lebesgue Measure on ℝⁿ: σ-Finiteness, Bounded Sets, Open Sets and Coordinate Hyperplanes (i)), so λm⊗λn\lambda_{m}\otimes\lambda_{n} is a measure on B(Rm)⊗B(Rn)\mathcal{B}(\mathbb{R}^{m})\otimes\mathcal{B}(\mathbb{R}^{n}) with

(λm⊗λn)(A×B)=λm(A) λn(B)(A∈B(Rm), B∈B(Rn))(\lambda_{m}\otimes\lambda_{n})(A\times B)=\lambda_{m}(A)\,\lambda_{n}(B)\qquad(A\in\mathcal{B}(\mathbb{R}^{m}),\ B\in\mathcal{B}(\mathbb{R}^{n}))

(Existence and Uniqueness of the Product Measure (i)). Since ι\iota is B(Rm)⊗B(Rn)/B(Rm+n)\mathcal{B}(\mathbb{R}^{m})\otimes\mathcal{B}(\mathbb{R}^{n})/\mathcal{B}(\mathbb{R}^{m+n})-measurable (Borel Sets of ℝ^{m+n} and the Product σ-Algebra of ℝ^m and ℝ^n (i)), ρ=(λm⊗λn)∘ι−1\rho=(\lambda_{m}\otimes\lambda_{n})\circ\iota^{-1} is a measure on (Rm+n,B(Rm+n))(\mathbb{R}^{m+n},\mathcal{B}(\mathbb{R}^{m+n})) (Image Measures and Measures with a Density: Measures, Integrals and Change of Variables (i)).

Let a,b∈Rm+na,b\in\mathbb{R}^{m+n} with ak<bka_{k}<b_{k} for every k∈[m+n]k\in[m+n], and let (a′,a′′)(a',a'') and (b′,b′′)(b',b'') be the pairs that ι\iota maps to aa and bb (Concatenation Identifies a Product of Euclidean Spaces with a Euclidean Space (iii)): ak′=aka'_{k}=a_{k} and bk′=bkb'_{k}=b_{k} for k∈[m]k\in[m], and aj′′=am+ja''_{j}=a_{m+j} and bj′′=bm+jb''_{j}=b_{m+j} for j∈[n]j\in[n]. By Open Boxes: Emptiness, Intersections and Preimages under Concatenation (iii),

ι−1(R(a,b))=R(a′,b′)×R(a′′,b′′),\iota^{-1}\bigl(R(a,b)\bigr)=R(a',b')\times R(a'',b''),

where R(a′,b′)∈B(Rm)R(a',b')\in\mathcal{B}(\mathbb{R}^{m}) and R(a′′,b′′)∈B(Rn)R(a'',b'')\in\mathcal{B}(\mathbb{R}^{n}), open boxes being Borel (Existence and Uniqueness of Lebesgue Measure on ℝⁿ (i), Closure Properties of σ-Algebras, Intersections and Traces of σ-Algebras, and the Generated σ-Algebra (iii)). The product formula therefore applies, and

ρ(R(a,b))=λm(R(a′,b′)) λn(R(a′′,b′′))=(∏k=1m(bk−ak))(∏j=1n(bm+j−am+j))=∏k=1m+n(bk−ak),\rho\bigl(R(a,b)\bigr)=\lambda_{m}\bigl(R(a',b')\bigr)\,\lambda_{n}\bigl(R(a'',b'')\bigr)=\Bigl(\prod_{k=1}^{m}(b_{k}-a_{k})\Bigr)\Bigl(\prod_{j=1}^{n}(b_{m+j}-a_{m+j})\Bigr)=\prod_{k=1}^{m+n}(b_{k}-a_{k}),

the last equality by Iterated Operations: Recursion, Splitting, Reordering, Termwise Combination and Homomorphisms (iii). Therefore ρ=λm+n\rho=\lambda_{m+n}. This proves (ii).

Images of product-measurable sets. For E∈B(Rm)⊗B(Rn)E\in\mathcal{B}(\mathbb{R}^{m})\otimes\mathcal{B}(\mathbb{R}^{n}), ι[E]∈B(Rm+n)\iota[E]\in\mathcal{B}(\mathbb{R}^{m+n}) by Borel Sets of ℝ^{m+n} and the Product σ-Algebra of ℝ^m and ℝ^n (ii), and ι−1(ι[E])=E\iota^{-1}(\iota[E])=E because ι\iota is a bijection (Images and Preimages under a Bijection and Its Inverse (ii)); so, by (ii),

λm+n(ι[E])=ρ(ι[E])=(λm⊗λn)(E).\lambda_{m+n}(\iota[E])=\rho(\iota[E])=(\lambda_{m}\otimes\lambda_{n})(E).

This proves (iii).

Nonnegative functions. Let f∈M+(Rm+n,B(Rm+n))f\in\mathcal{M}^{+}(\mathbb{R}^{m+n},\mathcal{B}(\mathbb{R}^{m+n})) and F=f∘ιF=f\circ\iota, so F(ξ,η)=f(ι(ξ,η))F(\xi,\eta)=f(\iota(\xi,\eta)). By Image Measures and Measures with a Density: Measures, Integrals and Change of Variables (ii) and (ii), F∈M+(Rm×Rn,B(Rm)⊗B(Rn))F\in\mathcal{M}^{+}(\mathbb{R}^{m}\times\mathbb{R}^{n},\mathcal{B}(\mathbb{R}^{m})\otimes\mathcal{B}(\mathbb{R}^{n})) and

∫Rm+nf dλm+n=∫Rm+nf d((λm⊗λn)∘ι−1)=∫Rm×RnF d(λm⊗λn).\int_{\mathbb{R}^{m+n}}f\,d\lambda_{m+n}=\int_{\mathbb{R}^{m+n}}f\,d\bigl((\lambda_{m}\otimes\lambda_{n})\circ\iota^{-1}\bigr)=\int_{\mathbb{R}^{m}\times\mathbb{R}^{n}}F\,d(\lambda_{m}\otimes\lambda_{n}).

Tonelli's theorem Tonelli's and Fubini's Theorems (i), applied to FF and the σ\sigma-finite measures λm\lambda_{m} and λn\lambda_{n}, says that each function η↦F(ξ,η)\eta\mapsto F(\xi,\eta) lies in M+(Rn,B(Rn))\mathcal{M}^{+}(\mathbb{R}^{n},\mathcal{B}(\mathbb{R}^{n})), that ξ↦∫RnF(ξ,η) dλn(η)\xi\mapsto\int_{\mathbb{R}^{n}}F(\xi,\eta)\,d\lambda_{n}(\eta) lies in M+(Rm,B(Rm))\mathcal{M}^{+}(\mathbb{R}^{m},\mathcal{B}(\mathbb{R}^{m})), and that the last integral above equals ∫Rm(∫RnF(ξ,η) dλn(η))dλm(ξ)\int_{\mathbb{R}^{m}}\bigl(\int_{\mathbb{R}^{n}}F(\xi,\eta)\,d\lambda_{n}(\eta)\bigr)d\lambda_{m}(\xi). This is (iv).

Integrable functions. Let f∈L1(λm+n)f\in\mathcal{L}^{1}(\lambda_{m+n}) and F=f∘ιF=f\circ\iota. By Image Measures and Measures with a Density: Measures, Integrals and Change of Variables (iii) and (ii), F∈L1(λm⊗λn)F\in\mathcal{L}^{1}(\lambda_{m}\otimes\lambda_{n}) and

∫Rm+nf dλm+n=∫Rm×RnF d(λm⊗λn).\int_{\mathbb{R}^{m+n}}f\,d\lambda_{m+n}=\int_{\mathbb{R}^{m}\times\mathbb{R}^{n}}F\,d(\lambda_{m}\otimes\lambda_{n}).

The section of FF at ξ∈Rm\xi\in\mathbb{R}^{m} is the function η↦f(ι(ξ,η))\eta\mapsto f(\iota(\xi,\eta)) (Precomposition Commutes with Pointwise Operations, and Sections of Functions Are Precompositions (iv)), so the set of ξ\xi at which this section does not lie in L1(λn)\mathcal{L}^{1}(\lambda_{n}) is the set NN of (v), and the function II of Tonelli's and Fubini's Theorems (ii) for FF is the function gg of (v). Fubini's theorem Tonelli's and Fubini's Theorems (ii), applied to FF and the σ\sigma-finite measures λm\lambda_{m} and λn\lambda_{n}, therefore gives N∈B(Rm)N\in\mathcal{B}(\mathbb{R}^{m}), λm(N)=0\lambda_{m}(N)=0 and g∈L1(λm)g\in\mathcal{L}^{1}(\lambda_{m}), with

∫Rm×RnF d(λm⊗λn)=∫Rmg dλm.\int_{\mathbb{R}^{m}\times\mathbb{R}^{n}}F\,d(\lambda_{m}\otimes\lambda_{n})=\int_{\mathbb{R}^{m}}g\,d\lambda_{m}.

This proves (v).

Citations

Loading…

Dependencies

Uses0

Loading…

Comments

Log in to comment.

Loading…