Skip to content

Substitution formula for one-demensional Lebesgue integral #1892

@Yosuke-Ito-345

Description

@Yosuke-Ito-345

@affeldt-aist @t6s @IshiguroYoshihiro

As I mentioned here, I am formalizing a substitution (change-of-variable) formula for Lebesgue integral. Here is the blueprint.

Theorem

Let $a, b \in \mathbb{R}$ with $a \leq b$. Let $f : [a, b] \to \mathbb{R}$ be of class $C^1$, and let $g : \mathbb{R} \to \mathbb{R}$ be a bounded Borel measurable function. Then

$$(g \circ f) \cdot f' : [a, b] \to \mathbb{R}; \quad t \mapsto g(f(t)) \cdot f'(t)$$

is Lebesgue integrable, and the following identity holds:

$$\int_{f(a)}^{f(b)} g(x) \, dx = \int_a^b g(f(t)) f'(t) \, dt.$$

Here both sides are oriented Lebesgue integrals (i.e. sign is taken into account).

Proof

Since $g$ is bounded, $g \circ f$ is also bounded. Moreover, since $f$ is $C^1$ on $[a, b]$, $f'$ is continuous on $[a, b]$, hence in particular bounded and Borel measurable. Therefore $(g \circ f) \cdot f'$ is Lebesgue integrable on $[a, b]$.

Define functions $F, G : [a, b] \to \mathbb{R}$ by

$$F(s) := \int_a^s g(f(t)) f'(t) \, dt,$$ $$G(s) := \int_{f(a)}^{f(s)} g(x) \, dx.$$

By the Fundamental Theorem of Calculus lemma FTC1,

$$F'(t) = g(f(t)) f'(t) \quad \text{for a.e.\ } t \in [a, b].$$

Furthermore, viewing $G$ as the composition of $s \mapsto f(s)$ and $r \mapsto \int_{f(a)}^{r} g$, an application of lemma FTC1 (with the chain rule for the derivative of the composite functions) gives

$$G'(s) = g(f(s)) f'(s) \quad \text{for a.e.\ } s \in [a, b].$$

Hence $F'(s) = G'(s)$ for a.e. $s \in [a, b]$. Moreover,

$$F(a) = G(a) = 0.$$

By lemma ae_eq_integral we have

$$F = G \quad \text{on\ } [a, b].$$

In particular, evaluating at $s = b$ gives $F(b) = G(b)$, i.e.,

$$\int_a^b g(f(t))\, f'(t)\, dt \;=\; \int_{f(a)}^{f(b)} g(x)\, dx.$$

Observation

I think the bottleneck is to introduce the oriented Lebesgue integral, which we have yet to formalize.

Disclaimer

The idea of the proof is inspired by Claude and ChatGPT, but I wrote down the proof and checked its accuracy by myself. Therefore, I am responsible for any trouble coming from this formalization.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions