Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Appendix A.6: clustering constraint supports lie in decomposition bags

Proved
StableMSR.Recovery.clustering_constraint_bag_placement

by lreyzin · Nov 13, 2026 · Mathlib 0df444a (Lean v4.33.1)

clusteringgraph-theorytreewidth

Let ddd be a finite metric, let R≥0R\ge0R≥0, and let each candidate set CiC_iCi​ lie within distance RRR of its specified center cic_ici​. Assign it to an owner vi∈Nv_i\in Nvi​∈N with d(vi,ci)≤2Rd(v_i,c_i)\le2Rd(vi​,ci​)≤2R. The interaction graph on NNN joins distinct vertices at distance at most 8R8R8R. Fix any finite tree decomposition of that graph.

Every point constraint fits in a single bag: for each input point xxx, some bag contains all owners of candidates covering xxx.

Every optional-component constraint also fits in a single bag: if an optional set KKK has diameter at most 2R2R2R, some bag contains all owners of candidates meeting KKK.

The result holds for every candidate family passing these geometric conditions, independently of stability or optimality. In the source algorithm, optional components satisfy the stronger diameter bound diam⁡(K)<2b<2R\operatorname{diam}(K)<2b<2Rdiam(K)<2b<2R. Thus this establishes the geometric support-containment hypothesis used to allocate the actual point predicates, optional-component predicates, and optional-component count/cost contributions to bags. It assumes a tree decomposition as input; constructing that decomposition and compiling the DP are separate obligations.

Preamble
import Definitions.Def_StableMSR_Recovery_Geometry
import Definitions.Def_RobertsonSeymour1986_GM5_IsTreeDecomposition

set_option autoImplicit false
open scoped Classical
Formal statement
namespace StableMSR.Recovery

theorem clustering_constraint_bag_placement {n : ℕ} {I T : Type}
    [Fintype T] [DecidableEq T]
    (d : Point n → Point n → ℝ) (hd : IsMetricArray d) (R : ℝ) (hR : 0 ≤ R)
    (N : Cluster n) (pool : Finset I) (candidate : I → Cluster n)
    (center : I → Point n) (owner : I → N)
    (howner : ∀ i ∈ pool, d (owner i) (center i) ≤ 2 * R)
    (hcandidate : ∀ i ∈ pool, ∀ x ∈ candidate i, d (center i) x ≤ R)
    (optional : Partition n) (hdiam : ∀ K ∈ optional, diameter d K ≤ 2 * R)
    (tree : SimpleGraph T) (bags : T → Finset N)
    (htd : RobertsonSeymour1986.GM5.IsTreeDecomposition
      ((thresholdGraph d (8 * R) N).induce (N : Set (Point n))) tree bags) :
    (∀ x : Point n, ∃ t : T, ∀ i ∈ pool, x ∈ candidate i → owner i ∈ bags t) ∧
      (∀ K ∈ optional, ∃ t : T, ∀ i ∈ pool,
        (candidate i ∩ K).Nonempty → owner i ∈ bags t) := by sorry

end StableMSR.Recovery
Source
Ravichandran and Reyzin, Exact Sum-of-Radii Clustering under Any Constant Stability, https://arxiv.org/html/2611.03547v1, Appendix A.6, paragraphs beginning "On every input, each constraint involves a clique" and "A tree decomposition is a tree of bags". The proof uses 3R owner-to-member bounds, 6R point-scope diameter, and 6R+diam(K)≤8R optional-scope diameter.

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me