Refining a partition increases the flip metric
Lax185574.FlipMetricMonotone · concepts/Lax185574/FlipMetricMonotone.lean · lax-185574
No public endorsements yet.
Loading review…
Sign in with ORCIDNatural Language Statement
Observation
If 𝒬 refines 𝒫, then every 𝒫-flip is also a 𝒬-flip, and hence
dist_𝒬(u, v) ≥ dist_𝒫(u, v).
Concept map
Evidence
In the paper
- page 27 of this submission's paper
Lean source view on GitHub
| 1 | import Lax185574.Flips |
| 2 | |
| 3 | /-! |
| 4 | --- |
| 5 | title: Refining a partition increases the flip metric |
| 6 | type: observation |
| 7 | --- |
| 8 | If 𝒬 refines 𝒫, then every 𝒫-flip is also a 𝒬-flip, and hence |
| 9 | |
| 10 | dist_𝒬(u, v) ≥ dist_𝒫(u, v). |
| 11 | |
| 12 | # Formalization notes |
| 13 | |
| 14 | Refinement is 𝒬 ≤ 𝒫 for setoids. The inequality of the flip metrics is |
| 15 | stated through their balls: for every r, Ball^𝒬_r(v) ⊆ Ball^𝒫_r(v), |
| 16 | that is, dist_𝒬(u, v) ≤ r implies dist_𝒫(u, v) ≤ r. This is the |
| 17 | inequality for distances that may be infinite. The two parts of the |
| 18 | observation are two axioms. The graph may be infinite. |
| 19 | -/ |
| 20 | |
| 21 | namespace Lax185574.FlipMetricMonotone |
| 22 | |
| 23 | open Lax185574.Flips |
| 24 | |
| 25 | universe u |
| 26 | |
| 27 | /-- If `𝒬` refines `𝒫`, every `𝒫`-flip is a `𝒬`-flip. -/ |
| 28 | axiom isFlip_of_le {V : Type u} (G : SimpleGraph V) {P Q : Setoid V} |
| 29 | (hQP : Q ≤ P) (H : SimpleGraph V) (hH : IsFlip G P H) : IsFlip G Q H |
| 30 | |
| 31 | /-- If `𝒬` refines `𝒫`, then `dist_𝒬(u, v) ≥ dist_𝒫(u, v)`: every ball |
| 32 | of the flip metric of `𝒬` lies in the ball of the same radius of the |
| 33 | flip metric of `𝒫`. -/ |
| 34 | axiom flipBall_subset_of_le {V : Type u} (G : SimpleGraph V) {P Q : Setoid V} |
| 35 | (hQP : Q ≤ P) (r : ℕ) (v : V) : flipBall G Q r v ⊆ flipBall G P r v |
| 36 | |
| 37 | end Lax185574.FlipMetricMonotone |
| 38 |
Formalization notes
Refinement is 𝒬 ≤ 𝒫 for setoids. The inequality of the flip metrics is stated through their balls: for every r, Ball^𝒬_r(v) ⊆ Ball^𝒫_r(v), that is, dist_𝒬(u, v) ≤ r implies dist_𝒫(u, v) ≤ r. This is the inequality for distances that may be infinite. The two parts of the observation are two axioms. The graph may be infinite.
Discussion
Ask a question or add context. Endorsements and structured flags are kept in the review panel above.
0 comments