Bourbaki

Corollary

Tag 08TO, permanent. Cite it as [Bourbaki, Tag 08TO].

Integration, chapter IV, § 4. Integrable functions and sets

— Let X be a locally compact space with a countable base, Ψ\Psi the set of Borel sets of X, β\beta a mapping of Ψ\Psi into [0,+∞][0, +\infty] satisfying the following conditions:

(i) If (B1,B2,…)(B_1, B_2, \ldots) is a sequence of pairwise disjoint Borel sets of X, then β(B1∪B2∪…)=β(B1)+β(B2)+…\beta(B_1 \cup B_2 \cup \ldots) = \beta(B_1) + \beta(B_2) + \ldots.

(ii) If B is a compact subset of X, then β(B)<+∞\beta(B) < +\infty.

Then, there exists one and only one positive measure μ\mu on X such that β(B)=μ∗(B)\beta(B) = \mu^*(B) for all B∈ΨB \in \Psi.

Let Φ\Phi be the set of compact subsets of XX and let α\alpha be the restriction of β\beta to Φ\Phi. The conditions (PCI_I), (PCII_{II}), (PMI_I), (PMII_{II}), (PMIII_{III}) and (PMIV′_{IV}') are then satisfied. Let KK be a compact subset of XX, and ε>0\varepsilon > 0. Then KK is the intersection of a decreasing sequence (U1,U2,…)(U_1, U_2, \ldots) of relatively compact open sets of XX (GT, IX, §2, No. 5, Prop. 7). We have ∑n=1∞β(Un−Un+1)=β(U1−K)<+∞\sum_{n=1}^{\infty} \beta(U_n - U_{n+1}) = \beta(U_1 - K) < +\infty, therefore

β(Un)−β(K)=β(Un−K)=∑p=n∞β(Up−Up+1) \beta(U_n) - \beta(K) = \beta(U_n - K) = \sum_{p=n}^{\infty} \beta(U_p - U_{p+1})

tends to 0 as nn tends to ∞\infty. This proves that the condition (PMIV′)(\mathrm{PM}'_{\mathrm{IV}}) is satisfied. By Th. 5, there exists a positive measure μ\mu on XX such that μ(K)=α(K)\mu(K) = \alpha(K) for every compact subset KK of XX. Since every open set UU of XX is the union of an increasing sequence of compact subsets, we have μ∗(U)=β(U)\mu^*(U) = \beta(U). Let LL be a compact subset of XX. By Prop. 7 of No. 5, the μ\mu-integrable subsets of LL form a tribe of subsets of LL. Therefore, if BB is an element of Ψ\Psi contained in LL, then BB is μ\mu-integrable; for every ε>0\varepsilon > 0, there then exist a compact set KK and an open set UU in XX such that K⊂B⊂UK \subset B \subset U and μ∗(U)−μ(K)≤ε\mu^*(U) - \mu(K) \leq \varepsilon (No. 6, Th. 4). Since β(U)=μ∗(U)\beta(U) = \mu^*(U) and β(K)=μ(K)\beta(K) = \mu(K), we see that ∣μ∗(B)−β(B)∣≤2ε|\mu^*(B) - \beta(B)| \leq 2\varepsilon. Therefore β(B)=μ∗(B)\beta(B) = \mu^*(B). Finally, every Borel set CC of XX is the union of a sequence of pairwise disjoint, relatively compact Borel sets, whence β(C)=μ∗(C)\beta(C) = \mu^*(C). The uniqueness of μ\mu follows at once from Th. 5.

Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.

Cited by

Nothing.

Cites

Transcribed from the printed volume. Not translated.