subadditive_f_generator_deriv2
For s continuous on (0, ∞) and y > 0, the derivative of the generator f of equation (4) has derivative α/y + s(y)/y² at y, i.e. f''(y) = α/y + s(y)/y², so y² f''(y) = α y + s(y).
Statement
theorem subadditive_f_generator_deriv2 (α : ℝ) (s : ℝ → ℝ)
(hs : ContinuousOn s (Set.Ioi 0)) {y : ℝ} (hy : 0 < y) :
HasDerivAt (deriv (subadditive_f_generator α s)) (α / y + s y / y ^ 2) y- Source
- helper for arXiv:2610.21231 (On the subadditivity of mutual f-information), Corollary 1 (equation (4) integrates x² f''(x) = g(x) = α x + s(x))
- Verified
- 23 Oct 2026
- Axioms
Classical.choiceQuot.soundpropext- Built on 2
- subadditive_f_generatorThe generator function f associated to parameter α and function s is f(x) = α x log x + ∫_1^x (x - t) (s(t) / t^2) dt.
- subadditive_f_generator_hasDerivAtFor s continuous on (0, ∞) and y > 0, the generator f(x) = α x log x + ∫_1^x (x - t) s(t)/t² dt has derivative f'(y) = α (log y + 1) +…
- Used by 2
- subadditive_f_generator_weight_eventuallyEqFor s continuous on (0, ∞) and x > 0, the weight g(y) = y² f''(y) of the generator f of equation (4) agrees with α y + s(y) on a…
- subadditive_f_generator_satisfies_weight_boundFirst claim of Corollary 1 of arXiv:2610.21231: let s be twice continuously differentiable and positive on (0, ∞) with s'' < 0, let K_s(x)…
The proof, explained
The proof computes the second derivative of the generator $f$ at $y > 0$ by differentiating the explicit formula for $f'$ established in an earlier lemma. On the open neighborhood $(0, \infty)$ of $y$, $f'(x)$ is given by $\alpha(\ln x + 1) + \int_1^x \frac{s(t)}{t^2}\,dt$. Differentiating the logarithmic term yields $\alpha / y$, while applying the Fundamental Theorem of Calculus to the integral term yields $s(y)/y^2$. Because $f'$ agrees with this sum throughout a neighborhood of $y$, its derivative at $y$ is the sum of the two derivatives, $\alpha/y + s(y)/y^2$.
The interval $(0, \infty)$ is open, and the integrand $t \mapsto s(t)/t^2$ is continuous on $(0, \infty)$.
The half-line $(0, \infty)$ is open by `isOpen_Ioi`. Because $s$ is continuous on $(0, \infty)$ and $t \mapsto t^2$ is continuous, their quotient $t \mapsto s(t)/t^2$ is continuous on $(0, \infty)$ by `ContinuousOn.div`, since $t > 0$ implies $t^2 > 0$ (`pow_pos`) and hence $t^2 \neq 0$ (`ne_of_gt`).
uses
isOpen_Ioi,ContinuousOn.div,Set.mem_Ioi,ne_of_gt,pow_poshave ho : IsOpen (Set.Ioi (0 : ℝ)) := isOpen_Ioi have hcs1 : ContinuousOn (fun t => s t / t ^ 2) (Set.Ioi 0) := by apply ContinuousOn.div hs (continuousOn_id.pow 2) intro t ht; rw [Set.mem_Ioi] at ht; exact ne_of_gt (pow_pos ht 2)
For any function $f$ continuous on $(0, \infty)$, the integral $u \mapsto \int_1^u f(t)\,dt$ has derivative $f(y)$ at $y$.
Let $f$ be continuous on $(0, \infty)$. Any $t$ in the unoriented interval $[1, y]$ lies between $1$ and $y$, so $t > 0$ because $y > 0$, showing $[1, y] \subseteq (0, \infty)$. Thus $f$ is continuous on $[1, y]$ and hence interval-integrable. Since $(0, \infty)$ is open, $f$ is continuous at each point of $(0, \infty)$, which implies that $f$ is strongly measurable with respect to the neighborhood filter of $y$ (`ContinuousAt.stronglyMeasurableAtFilter`) and continuous at $y$. By the Fundamental Theorem of Calculus (`intervalIntegral.integral_hasDerivAt_right`), the derivative of $u \mapsto \int_1^u f(t)\,dt$ at $y$ is $f(y)$.
uses
Set.mem_Ioi,Set.mem_uIcc,ContinuousAt.stronglyMeasurableAtFilter,intervalIntegral.integral_hasDerivAt_righthave h_deriv_integral : ∀ (f : ℝ → ℝ), ContinuousOn f (Set.Ioi 0) → HasDerivAt (fun u => ∫ t in (1 : ℝ)..u, f t) (f y) y := by intro f hf have hsub : Set.uIcc 1 y ⊆ Set.Ioi 0 := by intro t ht rw [Set.mem_Ioi] rw [Set.mem_uIcc] at ht rcases ht with ⟨h1, h2⟩ | ⟨h1, h2⟩ <;> linarith have hint : IntervalIntegrable f MeasureTheory.volume 1 y := (hf.mono hsub).intervalIntegrable have hfc : ∀ x ∈ Set.Ioi (0 : ℝ), ContinuousAt f x := fun x hx => hf.continuousAt (ho.mem_nhds hx) have hmeas : StronglyMeasurableAtFilter f (nhds y) MeasureTheory.volume := ContinuousAt.stronglyMeasurableAtFilter ho hfc y hy have hconty : ContinuousAt f y := hfc y hy exact intervalIntegral.integral_hasDerivAt_right hint hmeas hconty
The integral $x \mapsto \int_1^x s(t)/t^2\,dt$ has derivative $s(y)/y^2$ at $y$.
This is the specialization of the previous integral-derivative lemma to the function $t \mapsto s(t)/t^2$, which was shown to be continuous on $(0, \infty)$.
have hf1 : HasDerivAt (fun x => ∫ t in (1 : ℝ)..x, s t / t ^ 2) (s y / y ^ 2) y := h_deriv_integral (fun t => s t / t ^ 2) hcs1
The term $x \mapsto \alpha(\ln x + 1)$ has derivative $\alpha / y$ at $y$.
Because $y > 0$, we have $y \neq 0$ (`ne_of_gt`), so the natural logarithm has derivative $y^{-1}$ at $y$ (`Real.hasDerivAt_log`). Adding the constant $1$ leaves the derivative unchanged, and scaling by $\alpha$ gives derivative $\alpha \cdot y^{-1} = \alpha / y$.
uses
ne_of_gt,Real.hasDerivAt_loghave hy0 : y ≠ 0 := ne_of_gt hy have hlog : HasDerivAt Real.log y⁻¹ y := Real.hasDerivAt_log hy0 have hlog1 : HasDerivAt (fun x => Real.log x + 1) y⁻¹ y := by simpa using hlog.add_const 1 have hα : HasDerivAt (fun x => α * (Real.log x + 1)) (α / y) y := by have h := hlog1.const_mul α have hr : α * y⁻¹ = α / y := by ring rwa [hr] at h
The sum $x \mapsto \alpha(\ln x + 1) + \int_1^x s(t)/t^2\,dt$ has derivative $\alpha/y + s(y)/y^2$ at $y$.
The derivative of a sum is the sum of the derivatives of the two terms, giving $\alpha/y + s(y)/y^2$.
have hsum : HasDerivAt (fun x => α * (Real.log x + 1) + ∫ t in (1 : ℝ)..x, s t / t ^ 2) (α / y + s y / y ^ 2) y := hα.add hf1
The derivative of the generator, $f'$, has derivative $\alpha/y + s(y)/y^2$ at $y$.
By `subadditive_f_generator_hasDerivAt`, for all $x \in (0, \infty)$, the derivative $f'(x)$ equals $\alpha(\ln x + 1) + \int_1^x s(t)/t^2\,dt$. Because $(0, \infty)$ is an open neighborhood of $y$, this equality holds in the neighborhood filter of $y$ (`Filter.eventuallyEq_iff_exists_mem.mpr`). Since differentiability at a point depends only on the values in a neighborhood, $f'$ has the same derivative at $y$ as this sum, namely $\alpha/y + s(y)/y^2$.
uses
Filter.eventuallyEq_iff_exists_mem.mpr,subadditive_f_generator_hasDerivAthave heq : deriv (subadditive_f_generator α s) =ᶠ[nhds y] fun x => α * (Real.log x + 1) + ∫ t in (1 : ℝ)..x, s t / t ^ 2 := by apply Filter.eventuallyEq_iff_exists_mem.mpr refine ⟨Set.Ioi 0, ho.mem_nhds hy, fun x hx => ?_⟩ exact (subadditive_f_generator_hasDerivAt α s hs hx).deriv exact hsum.congr_of_eventuallyEq heq
ContinuousOn.div- The quotient of two functions continuous on a set $s$ is continuous on $s$, provided the denominator does not vanish on $s$.
ContinuousAt.stronglyMeasurableAtFilter- If a function is continuous on an open set $s$, then it is strongly measurable with respect to the neighborhood filter of each point in $s$.
Filter.eventuallyEq_iff_exists_mem.mpr- Two functions are eventually equal with respect to a filter if they agree on a set belonging to that filter.
Real.hasDerivAt_log- For any non-zero real number $x$, the natural logarithm function has derivative $x^{-1} = 1/x$ at $x$.
Set.mem_Ioi- A point $x$ belongs to the open interval $(b, \infty)$ if and only if $b < x$.
Set.mem_uIcc- A point belongs to the unoriented closed interval between $a$ and $b$ if and only if it lies between them.
intervalIntegral.integral_hasDerivAt_right- If $f$ is integrable on the interval between $a$ and $b$, strongly measurable near $b$, and continuous at $b$, then $u \mapsto \int_a^u f(t)\,dt$ has derivative $f(b)$ at $u = b$.
isOpen_Ioi- The open half-line $(a, \infty)$ is an open set.
ne_of_gt- If $b < a$, then $a \neq b$.
pow_pos- If $a > 0$, then $a^n > 0$ for any natural number $n$.
subadditive_f_generator_hasDerivAt- For $s$ continuous on $(0, \infty)$ and $y > 0$, the generator function $f$ has derivative $\alpha(\ln y + 1) + \int_1^y \frac{s(t)}{t^2}\,dt$ at $y$.
- In paper proofs, one typically differentiates the explicit formula for $f'(x)$ directly without mentioning filter-eventual equality. In Lean, because the theorem statement concerns $\text{deriv}(f)$, one must show that $\text{deriv}(f)$ agrees with the formula on a neighborhood of $y$ using `eventuallyEq` and `congr_of_eventuallyEq`.
- Applying the Fundamental Theorem of Calculus for Bochner integrals (`intervalIntegral.integral_hasDerivAt_right`) requires proving that the integrand is strongly measurable with respect to the neighborhood filter of $y$, which Mathlib deduces from continuity on the open set $(0, \infty)$.
Written by a model (gemini-3.8-flash) from the Lean proof and the goal states Lean reported at each step. Lean verified the proof; this explanation is a reading aid and is not itself machine-checked.
Lean source view module on GitHub
Choose a tactic — a marked line — to see the goals Lean had before and after it, as in Lean’s infoview. Step with ← → or j k.
theorem subadditive_f_generator_deriv2 (α : ℝ) (s : ℝ → ℝ) (hs : ContinuousOn s (Set.Ioi 0)) {y : ℝ} (hy : 0 < y) : HasDerivAt (deriv (subadditive_f_generator α s)) (α / y + s y / y ^ 2) y := by have ho : IsOpen (Set.Ioi (0 : ℝ)) := isOpen_Ioi have hcs1 : ContinuousOn (fun t => s t / t ^ 2) (Set.Ioi 0) := by apply ContinuousOn.div hs (continuousOn_id.pow 2) intro t ht; rw [Set.mem_Ioi] at ht; exact ne_of_gt (pow_pos ht 2) have h_deriv_integral : ∀ (f : ℝ → ℝ), ContinuousOn f (Set.Ioi 0) → HasDerivAt (fun u => ∫ t in (1 : ℝ)..u, f t) (f y) y := by intro f hf have hsub : Set.uIcc 1 y ⊆ Set.Ioi 0 := by intro t ht rw [Set.mem_Ioi] rw [Set.mem_uIcc] at ht rcases ht with ⟨h1, h2⟩ | ⟨h1, h2⟩ <;> linarith have hint : IntervalIntegrable f MeasureTheory.volume 1 y := (hf.mono hsub).intervalIntegrable have hfc : ∀ x ∈ Set.Ioi (0 : ℝ), ContinuousAt f x := fun x hx => hf.continuousAt (ho.mem_nhds hx) have hmeas : StronglyMeasurableAtFilter f (nhds y) MeasureTheory.volume := ContinuousAt.stronglyMeasurableAtFilter ho hfc y hy have hconty : ContinuousAt f y := hfc y hy exact intervalIntegral.integral_hasDerivAt_right hint hmeas hconty have hf1 : HasDerivAt (fun x => ∫ t in (1 : ℝ)..x, s t / t ^ 2) (s y / y ^ 2) y := h_deriv_integral (fun t => s t / t ^ 2) hcs1 have hy0 : y ≠ 0 := ne_of_gt hy have hlog : HasDerivAt Real.log y⁻¹ y := Real.hasDerivAt_log hy0 have hlog1 : HasDerivAt (fun x => Real.log x + 1) y⁻¹ y := by simpa using hlog.add_const 1 have hα : HasDerivAt (fun x => α * (Real.log x + 1)) (α / y) y := by have h := hlog1.const_mul α have hr : α * y⁻¹ = α / y := by ring rwa [hr] at h have hsum : HasDerivAt (fun x => α * (Real.log x + 1) + ∫ t in (1 : ℝ)..x, s t / t ^ 2) (α / y + s y / y ^ 2) y := hα.add hf1 have heq : deriv (subadditive_f_generator α s) =ᶠ[nhds y] fun x => α * (Real.log x + 1) + ∫ t in (1 : ℝ)..x, s t / t ^ 2 := by apply Filter.eventuallyEq_iff_exists_mem.mpr refine ⟨Set.Ioi 0, ho.mem_nhds hy, fun x hx => ?_⟩ exact (subadditive_f_generator_hasDerivAt α s hs hx).deriv exact hsum.congr_of_eventuallyEq heq