graph_lly_curvature_eq_of_ollivier_eq
If κ_α(x, y) = (1 − α)c for all α in an interval (a, 1), then κ_LLY(x, y) = c.
Statement
theorem graph_lly_curvature_eq_of_ollivier_eq {V : Type*} [Fintype V] [DecidableEq V] (G : SimpleGraph V) [DecidableRel G.Adj] (x y : V) (c a : ℝ) (ha : a < 1) (h : ∀ α, a < α → α < 1 → graph_ollivier_curvature G α x y = (1 - α) * c) : graph_lly_curvature G x y = c
- Source
- Helper for the refutation of OP-165 (arXiv:2610.10559, after Theorem 1.4).
- Verified
- 10 Oct 2026
- Axioms
Classical.choiceQuot.soundpropext- Built on 2
- graph_lly_curvatureThe Lin–Lu–Yau curvature κ_LLY(x, y) = lim_{α → 1⁻} κ_α(x, y)/(1 − α).
- graph_ollivier_curvatureThe α-Ollivier–Ricci curvature κ_α(x, y) = 1 − W(μ_x^α, μ_y^α)/d(x, y).
- Used by 4
- graph_lly_curvature_of_twinsIf x and y are adjacent and have the same neighbours apart from each other, then κ_LLY(x, y) = (d_x + 1)/d_x.
- not_edge_connectivity_ge_half_scale_mul_min_degree_mul_lly_lower_boundThe inequality κ'(G) ≥ (d/2) δ(G) κ_LLY^(d)(G) asked in OP-164 fails for d = 4: take a 4-cycle 0–1–2–3–0 and a triangle 4–5–6 joined by…
- graph_lly_curvature_of_leafIf x is a leaf of a finite connected graph whose only neighbour is y, then κ_LLY(x, y) = 2/d_y.
- lly_five_vertex_t_star_curvatureFor the tree on 5 vertices with edges 0–1, 0–3, 0–4, 1–2 one has κ_LLY(0,1) = −1/3.
The proof, explained
The Lin-Lu-Yau curvature κ_LLY(x, y) is defined as the left limit of κ_α(x, y) / (1 − α) as α approaches 1. Because the open interval (a, 1) is a left-neighborhood of 1, it suffices to show that this quotient is constantly c for all α ∈ (a, 1). Substituting the given Ollivier curvature hypothesis κ_α(x, y) = (1 − α)c, the non-zero factor 1 − α cancels out, leaving c identically on the interval.
To show κ_LLY(x, y) = c, it suffices to show that the quotient κ_α(x, y) / (1 − α) is eventually equal to c as α → 1⁻.
By definition, the Lin-Lu-Yau curvature κ_LLY(x, y) is the limit of κ_α(x, y) / (1 − α) along the left-neighborhood filter of 1. By Filter.Tendsto.limUnder_eq, this limit equals c if the quotient converges to c. Since the constant function c converges to c, tendsto_const_nhds.congr' reduces convergence to showing that the quotient is eventually equal to c along that filter.
uses
Filter.Tendsto.limUnder_eq,tendsto_const_nhds.congr'unfold graph_lly_curvature apply Filter.Tendsto.limUnder_eq apply tendsto_const_nhds.congr'
It suffices to show that κ_α(x, y) / (1 − α) = c for all α ∈ (a, 1).
Since a < 1, the open interval (a, 1) is a neighborhood of 1 from the left by Ioo_mem_nhdsLT. Therefore, the eventual equality holds if the quotient equals c everywhere on (a, 1).
uses
Ioo_mem_nhdsLTfilter_upwards [Ioo_mem_nhdsLT ha] with α hα
For every α ∈ (a, 1), the quotient κ_α(x, y) / (1 − α) equals c.
For α ∈ (a, 1), the hypothesis h substitutes (1 − α)c for κ_α(x, y). Since α < 1, the difference 1 − α is nonzero by sub_ne_zero.mpr. Finally, field_simp cancels the factor 1 − α from the numerator and denominator, reducing the expression to c.
uses
sub_ne_zero.mprrw [h α hα.1 hα.2] have : (1 - α) ≠ 0 := sub_ne_zero.mpr hα.2.ne' field_simp
Filter.Tendsto.limUnder_eq- If g converges to x along a filter f (with f ≠ ⊥ in a Hausdorff space), then the limit limUnder of g along f is equal to x.
tendsto_const_nhds.congr'- If a function is eventually equal to the constant c along a filter, then it converges to c along that filter.
Ioo_mem_nhdsLT- For a < b, the open interval (a, b) belongs to the punctured left-neighborhood filter of b.
sub_ne_zero.mpr- If a ≠ b, then a - b ≠ 0.
Written by a model (gemini-3.8-flash) from the Lean proof and the goal states Lean reported at each step. Lean verified the proof; this explanation is a reading aid and is not itself machine-checked.
Lean source view module on GitHub
Choose a tactic — a marked line — to see the goals Lean had before and after it, as in Lean’s infoview. Step with ← → or j k.
theorem graph_lly_curvature_eq_of_ollivier_eq {V : Type*} [Fintype V] [DecidableEq V] (G : SimpleGraph V) [DecidableRel G.Adj] (x y : V) (c a : ℝ) (ha : a < 1) (h : ∀ α, a < α → α < 1 → graph_ollivier_curvature G α x y = (1 - α) * c) : graph_lly_curvature G x y = c := by unfold graph_lly_curvature apply Filter.Tendsto.limUnder_eq apply tendsto_const_nhds.congr' filter_upwards [Ioo_mem_nhdsLT ha] with α hα rw [h α hα.1 hα.2] have : (1 - α) ≠ 0 := sub_ne_zero.mpr hα.2.ne' field_simp