Zulip Chat Archive

Stream: Is there code for X?

Topic: Discrete version of Dominated Convergence Theorem


Gareth Ma (Jul 20 2024 at 11:07):

Hi, does this version of the dominated convergence theorem exist in Mathlib? Basically if limNfN(n)=f(n)\lim_{N \to \infty} f_N(n) = f(n) then limNn=fN(n)=n=f(n)\lim_{N \to \infty} \sum_{n = -\infty}^{\infty} f_N(n) = \sum_{n = -\infty}^{\infty} f(n) (with some summability stuff of course)

Daniel Weber (Jul 20 2024 at 12:05):

Is this a valid formalization of the statement?

import Mathlib

open Filter Topology

example {α : Type*} {β : Type*} [NormedAddCommGroup α] (f_N : β    α) (g : β  α) (f : β  α)
    (dom :  x N, f_N x N  g x) (s : α) (h₂ : HasSum f s) (hg : Summable g)
    (h :  x, Tendsto (f_N x) atTop (𝓝 (f x))) : Tendsto (fun n  ∑' x, f_N x n) atTop (𝓝 s) := by
  sorry

Gareth Ma (Jul 20 2024 at 12:12):

I think so

Gareth Ma (Jul 21 2024 at 13:09):

If you have time, can you try to prove it by applying DCT with the counting measure? Or I think this is a simpler version:

example {α : Type*} {β : Type*} [NormedAddCommGroup α]
    (f_N : β    α) (g : β  α) (f : β  α)
    (dom :  x N, f_N x N  g x) (hg : Summable g) (h :  x, Tendsto (f_N x) atTop (𝓝 (f x))) :
    HasSum (fun n  ∑' x, f_N x n) (∑' x, f x) := by

I tried but I am not familiar with the measure theory / analysis part of Mathlib at all :(


Last updated: May 02 2025 at 03:31 UTC