Zulip Chat Archive

Stream: Is there code for X?

Topic: antitone function tendsto its liminf


Bhavik Mehta (Mar 19 2025 at 04:52):

Do we have anything like this?

import Mathlib

open ENNReal Filter Topology

lemma tendsto_liminf {f : 0  0} (hf : Antitone f) :
    Tendsto f (𝓝[>] 0) (𝓝 (liminf f (𝓝[>] 0))) := by
  sorry

I'd be equally happy with the limit expressed as limsup, or as ⨆ r, ⨆ (_ : a < r), f r, but everything I can find is where the filter is atTop, atBot, or nhds.

Bhavik Mehta (Mar 19 2025 at 04:52):

(deleted)

Bhavik Mehta (Mar 19 2025 at 04:52):

@loogle Filter.Tendsto, Filter.liminf, Antitone

loogle (Mar 19 2025 at 04:52):

:shrug: nothing found

Bhavik Mehta (Mar 19 2025 at 04:55):

My broader context is I have a goal of this shape:

example {f g : 0  0} {φ ψ :   0}
    (hf : Antitone f) (hg : Antitone g)
    ( : Antitone φ) (hφ₀ :  n, 0 < φ n) ( : Antitone ψ) (hψ₀ :  n, 0 < ψ n)
    (hφ_lim : Tendsto φ atTop (𝓝[>] 0)) (hψ_lim : Tendsto ψ atTop (𝓝[>] 0))
    (hfg :  n, f (φ n)  g (φ n) + ψ n) :
     r,  (_ : 0 < r), f r   r,  (_ : 0 < r), g r := by
  sorry

where the informal proof should be just "take n big in hfg", but I can't figure out a filter way of doing that nicely

Bhavik Mehta (Mar 19 2025 at 05:01):

(and in case it's useful, here's my lemma for passing between liminf and iSup2, which I have proved)

lemma iSup₂_eq_liminf {α β : Type*}
    [LinearOrder α] [TopologicalSpace α] [OrderTopology α] [DenselyOrdered α]
    [CompleteLattice β]
    (a : α) (hb :  b, a < b) {f : α  β} (hf : Antitone f) :
     r,  (_ : a < r), f r = Filter.liminf f (𝓝[>] a) := by

Bhavik Mehta (Mar 19 2025 at 05:11):

Ah! Found it, docs#Antitone.tendsto_nhdsGT was my missing ingredient


Last updated: May 02 2025 at 03:31 UTC