Zulip Chat Archive
Stream: Is there code for X?
Topic: Summability of g ∘ f
Michael Stoll (Nov 10 2023 at 21:10):
Do we have something like the following?
import Mathlib
lemma summable_comp_of_summable_and_differentiableAt {f : ℕ → ℂ} {g : ℂ → ℂ}
(hf : Summable fun x ↦ ‖f x‖) (hg : DifferentiableAt ℂ g 0) (h : g 0 = 0) :
Summable (g ∘ f) := sorry
(or possibly in a more general version)
I can't find anything. So this is perhaps a nice challenge for our topologists :smile:
EDIT: added missing condition g 0 = 0
.
Junyan Xu (Nov 10 2023 at 21:48):
Do you not need g 0 = 0
?
Patrick Massot (Nov 10 2023 at 21:51):
Like Junyan, I have a counter-example.
Junyan Xu (Nov 10 2023 at 21:51):
The general assumption should be g(y) = O(y)
as y->0, right
Yury G. Kudryashov (Nov 11 2023 at 02:55):
And we have docs#summable_of_isBigO_nat'
Yury G. Kudryashov (Nov 11 2023 at 02:57):
You can apply it to (hg : g =O[nhds 0] id).comp_tendsto (hf : Summable f).tendsto_atTop_zero
Michael Stoll (Nov 11 2023 at 08:19):
Junyan Xu said:
Do you not need
g 0 = 0
?
Yes, of course. I realized that after going to bed (it was late, and I was tired...). But I figured that whoever would attempt to prove it would realize that, too. I'm editing my first message to include it.
Michael Stoll (Nov 11 2023 at 12:18):
OK, thanks. What I have now is this.
open Topology in
lemma isBigO_id_of_differentiableAt {g : ℂ → ℂ} (hg : DifferentiableAt ℂ g 0) (hg₀ : g 0 = 0) :
g =O[𝓝 0] id := by
simp only [← hasDerivAt_deriv_iff, hasDerivAt_iff_isLittleO, hg₀, sub_zero, smul_eq_mul] at hg
have H := Asymptotics.IsBigO.const_smul_left
(Asymptotics.isBigO_refl (fun z : ℂ ↦ z) (𝓝 0)) <| deriv g 0
convert Asymptotics.IsBigO.add_isLittleO H hg using 1
simp only [Pi.smul_apply, smul_eq_mul, mul_comm (deriv _ _), add_sub_cancel'_right]
lemma summable_comp_of_summable_and_differentiableAt {f : ℕ → ℂ} {g : ℂ → ℂ}
(hf : Summable f) (hg : DifferentiableAt ℂ g 0) (hg₀ : g 0 = 0) :
Summable (g ∘ f) :=
summable_of_isBigO_nat' hf <|
(isBigO_id_of_differentiableAt hg hg₀).comp_tendsto hf.tendsto_atTop_zero
Is there a neater way to prove the first lemma? I'm just feeling around my way through this part of Mathlib...
Yury G. Kudryashov (Nov 11 2023 at 17:28):
Yury G. Kudryashov (Nov 11 2023 at 17:35):
I'm going to add IsBigO.comp_summable
in another PR.
Yury G. Kudryashov (Nov 12 2023 at 00:03):
Last updated: Dec 20 2023 at 11:08 UTC