Zulip Chat Archive
Stream: new members
Topic: Dealing with inequalities
Christian K (Mar 18 2024 at 16:46):
import Mathlib.Data.Real.Basic
import Mathlib.Data.Real.Sqrt
import Mathlib.Analysis.SpecialFunctions.Trigonometric.Basic
noncomputable abbrev sq_2 := Real.sqrt 2
example : -(1 / 2) ≤ 1 / 2 * Real.cos (↑n * sq_2) := by
sorry
example : Real.sin (↑n * sq_2) * Real.sin sq_2 * (-1 / 2) ≤
1 / 2 + Real.cos (↑n * sq_2) * Real.cos sq_2 * (1 / 2) := by
sorry
I want to prove these two inequalites, but I have no idea how to simplify them. With equalities I would use apply_fun to simplify the expression, but this does not work for <=. It says that I need to prove that the function I'm using is monotone but I have no idea how.
Do you have some general tips about proving such inequalities?
Jireh Loreaux (Mar 18 2024 at 16:48):
For the first you should use docs#Real.neg_one_le_cos
Christian K (Mar 18 2024 at 16:49):
Yeah, I know about this, but how do i get rid of the 1/2?
Jireh Loreaux (Mar 18 2024 at 16:52):
gcongr
is a handy tactic for this kind of thing:
example : -(1 / 2) ≤ 1 / 2 * Real.cos (↑n * sq_2) := by
calc
-(1 / 2) ≤ (1 / 2) * (-1) := by simp
_ ≤ (1 / 2) * (Real.cos (n * sq_2)) := by
gcongr
exact Real.neg_one_le_cos (n * sq_2)
Jireh Loreaux (Mar 18 2024 at 16:54):
Also, what is the type of n
supposed to be? With autoimplicits on, Lean assumes it to be ℝ
, but you have a ↑
in your statement. Is n
supposed to have type ℕ
?
Christian K (Mar 18 2024 at 16:57):
Ohh, i forgot to include this in the mwe, you guessed it correctly, n is supposed to have type Nat.
Christian K (Mar 18 2024 at 17:36):
And how about the second lemma, I don't really see how i should use a calc proof with the gcongr tactic to solve it?
Christian K (Mar 18 2024 at 17:36):
lemma sin_le_cos (x : Real) : Real.sin x ≤ 1 + Real.cos x := by
sorry
could a lemma like this help?
Jireh Loreaux (Mar 18 2024 at 19:12):
that inequality is not true.
Jireh Loreaux (Mar 18 2024 at 19:12):
Take x := -π/3
Jireh Loreaux (Mar 18 2024 at 19:15):
Your second result follows from the angle addition formula docs#Real.cos_sub
Christian K (Mar 19 2024 at 09:52):
Ohh ok, thank you, I figured it out
Last updated: May 02 2025 at 03:31 UTC