Lemma C.1 — hub construction and runner-up optimum
ProvedStableMSR.Recovery.Supplement.hub_runner_upclusteringp2o-batch-stoc27sp2o-gran-per-chapterp2o-plan-paperp2o-v1perturbation-stability
For a rational finite metric , , positive rational , and rational , form the hub/dummy instance of Appendix C. Let , , , , and . Its prescribed partition satisfies
The competing minimum is attained, the constructed distance array is metric, and is the unique original optimum. This pins down the entire construction and the exact runner-up value used in the hardness reduction.
Differences from the source:
- encoding: Finite ordered metric input and unordered exact-k partitions → Fin n points, real metric arrays, and finite sets of nonempty disjoint blocks; reason: Input-order relabelling preserves every source instance and partition; base published definitions are reused without extra conditions
- encoding: Minimum over different exact-K partitions in the constructed metric → An explicit competitor attaining delta+a, a lower bound for every competitor, and the metric property from the construction preamble; reason: This states the minimum with attainment rather than a possibly default real infimum; all parameters remain those of Appendix C
Preamble
import Definitions.Def_StableMSR_Recovery_Supplement_Hub set_option autoImplicit false open scoped Classical
Formal statement
namespace StableMSR.Recovery.Supplement
theorem hub_runner_up {n k : ℕ} (d : Point n → Point n → ℚ) (γ τ : ℚ)
(hk : 1 ≤ k) (hkn : k ≤ n) (hd : IsMetricArray (fun x y => (d x y : ℝ)))
(hγ : 1 < γ) (hτ : 0 < τ) :
let dr := fun x y => (d x y : ℝ)
let a := optimumValue dr k
let dp := hubDistance dr k γ τ
let O₀ := suppliedPartition n k
a ≤ diameter dr Finset.univ ∧ diameter dr Finset.univ < hubRadius dr γ τ ∧
IsMetricArray dp ∧ cost dp O₀ = hubRadius dr γ τ ∧
(∃ Q, Feasible Finset.univ (k + 2) Q ∧ Q ≠ O₀ ∧ cost dp Q = dummyDistance dr γ τ + a) ∧
(∀ Q, Feasible Finset.univ (k + 2) Q → Q ≠ O₀ → dummyDistance dr γ τ + a ≤ cost dp Q) ∧
UniqueOptimal dp Finset.univ (k + 2) O₀ := by sorry
end StableMSR.Recovery.Supplement
Source
Shyam Ravichandran, Lev Reyzin, STOC 2027 submission #1010 (October 2026), Lemma C.1, p. 24; PDF SHA256 8bb34b0b6d31458fe6fc0848c1fc82a6083f3e9ae672e748d24cf82d5289e53e