NBV 函数的导数与测度的关系

命题:μF⊥m  ⟺  F′=0a.e.\mu_F \perp m \iff F' = 0 \text{a.e.};μF≪m  ⟺  F(x)=∫−∞xF′\mu_F \ll m \iff F(x) = \int_{-\infty}^{x} F'

prop.nbv-derivative · 命题 · NBV · RN 导数

又称NBV 的导数与测度

图外条目Wikipedia ↗·nLab ↗

陈述

设 F∈NBVF \in NBV。则 F′∈L1(m)F' \in L^1(m),并且

μF⊥m  ⟺  F′=0a.e.\mu_F \perp m \iff F' = 0\quad \text{a.e.}μF≪m  ⟺  F(x)=∫−∞xF′(t)dt\mu_F \ll m \iff F(x) = \int_{-\infty}^{x} F'(t) dt

说明

证明:写 dμF=dλ+f⋅dmd\mu_F = d\lambda + f\cdot dm(Lebesgue 分解),由测度那条定理 f=F′f = F' a.e.。于是

⋅\cdot μF⊥m\mu_F \perp m 意味着绝对连续部为零,即 f=0  ⟹  F′=0a.e.f = 0 \implies F' = 0 \text{a.e.};

⋅\cdot μF≪m\mu_F \ll m 意味着 λ=0\lambda = 0,于是 F(x)=μF((−∞,x])=∫−∞xf⋅dm=∫−∞xF′⋅dtF(x) = \mu_F((-\infty,x]) = \int_{-\infty}^x f\cdot dm = \int_{-\infty}^x F'\cdot dt。

⭐ 这一条就是「微积分基本定理」的测度版:能不能把函数积回来,取决于 μF\mu _F 是否绝对连续。

English draft

Translation of the statement above · machine-drafted, pending author review · the Chinese text is authoritative.

Abstract

For FF of normalised bounded variation, F′∈L1(m)F' \in L^1(m) and μF⊥m  ⟺  F′=0\mu_F \perp m \iff F' = 0 a.e., while μF≪m  ⟺  F(x)=∫−∞xF′(t) dt\mu_F \ll m \iff F(x) = \int_{-\infty}^{x} F'(t)\,dt. Both follow by writing the Lebesgue decomposition dμF=dλ+f dmd\mu_F = d\lambda + f\,dm and using f=F′f = F' a.e.: vanishing of the absolutely continuous part is the first case, λ=0\lambda = 0 is the second. This is the measure-theoretic form of the fundamental theorem of calculus — F′F' integrates back to FF exactly when μF\mu_F is absolutely continuous.

它指向谁(出边)

推出 2

谁指向它(入边)

定义引用 5

参考

同一星团的其他概念 16 个

继续

在星图里打开「NBV 函数的导数与测度的关系」 ·  全部 709 个概念 ·  Markdown 原文 ·  English entry · 读图指南与记号表