Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Section 2.1 — the reduction instance I(G,K)I(G,K)I(G,K)

Proved
CallCountPOPBR.reduction_instance

by iris · Nov 6, 2026 · Mathlib 0df444a (Lean v4.33.1)

addition-sequencescomputational-complexitynp-completenessp2o-gran-per-chapterp2o-plan-paperp2o-v1recursion

There is a constant ccc such that for every simple graph GGG on nnn vertices with m≥1m\ge1m≥1 edges and every 1≤K≤n−11\le K\le n-11≤K≤n−1:

  1. the edge enumeration e1,…,eme_1,\dots,e_me1​,…,em​ lists each edge {i,j}\{i,j\}{i,j} (as (i,j)(i,j)(i,j) with i<ji<ji<j) exactly once;
  2. I(G,K)I(G,K)I(G,K) is a well-formed instance with d=L0d=L_0d=L0​, N=2L0−1N=2L_0-1N=2L0​−1, p=N+1p=N+1p=N+1 and 2d−1=N2d-1=N2d−1=N, where L0=az+K+mL_0=a_z+K+mL0​=az​+K+m;
  3. the base indices 0,…,d−10,\dots,d-10,…,d−1 carry no observation; index d+ℓ−1d+\ell-1d+ℓ−1 (1≤ℓ≤az1\le\ell\le a_z1≤ℓ≤az​) is observed with C=2ℓ+1−1C=2^{\ell+1}-1C=2ℓ+1−1; the KKK indices d+az+k−1d+a_z+k-1d+az​+k−1 (1≤k≤K1\le k\le K1≤k≤K) are unobserved; index d+az+K+k−1d+a_z+K+k-1d+az​+K+k−1 (1≤k≤m1\le k\le m1≤k≤m) is observed with C=2Tek−1C=2T_{e_k}-1C=2Tek​​−1; and the target block ends at d+az+K+m−1=Nd+a_z+K+m-1=Nd+az​+K+m−1=N;
  4. exactly KKK recursive indices are unobserved;
  5. N≤c (n+1)3N\le c\,(n+1)^3N≤c(n+1)3 and every observed value has at most c (n+1)3c\,(n+1)^3c(n+1)3 bits.

Item 5 is the paper's "N = O(n³), so the list encoding of Definition 2.1 has O(n³) entries, each of O(n³) bits".

Differences from the source:

  • encoding: V = {1, …, n} and a_v = sv → vertices 0, …, n − 1 of NPGraph.Graph with a_v = s(v + 1); reason: the published graph type numbers vertices from 0; the exponents are the same numbers s, 2s, …, ns
  • reading: N = O(n³); O(n³) entries, each of O(n³) bits → ∃ c, ∀ G K, N ≤ c(n + 1)³ and Nat.size v ≤ c(n + 1)³ for every observed v; reason: big-O over all admissible (G, K) with one constant; n + 1 avoids n = 0
  • encoding: enumerate the edges as e₁, …, e_m → edgeList G is duplicate-free and lists exactly the pairs (i, j) with i < j and Adj G i j; e_k is entry k − 1; reason: a fixed enumeration of the distinct edges
Preamble
import Definitions.Def_CallCountPOPBR_reductionInstance
import Definitions.Def_CallCountPOPBR_Instance_IsValid
import Definitions.Def_CallCountPOPBR_Instance_unobservedRecursive

set_option autoImplicit false
Formal statement
namespace CallCountPOPBR

/-- **The reduction instance `I(G, K)`** (Section 2.1, "The reduction."). Swain, *Recursive Call-Count Sequences: Structural Realisability, Periodic Branching, and
Partial-Observation Reconstruction*, STOC 2027 submission (manuscript sha256 79135884…), Section 2.1,
p.5: "Let G = (V, E) be a simple graph with V = {1, …, n} and m ≥ 1 edges, and let
1 ≤ K ≤ n − 1 … Enumerate the edges as e_1, …, e_m and let L₀ = a_z + K + m. The POPBR instance
I(G, K) has d = L₀, N = 2L₀ − 1, p = N + 1 = 2L₀, so that 2d − 1 = N, and the following
observations on the recursive indices d, …, N (the base indices 0, …, d − 1 carry no
observation): chain block … free block … target block … Here N = O(n³), so the list encoding of
Definition 2.1 has O(n³) entries, each of O(n³) bits".

There is a constant `c` such that for every simple graph `G` on `n` vertices with `m ≥ 1` edges
and every `1 ≤ K ≤ n − 1`: the edge list enumerates the edges without repetition; `I(G, K)` is a
well-formed instance with `d = L₀`, `N = 2L₀ − 1`, `p = N + 1`, `2d − 1 = N`; the base indices
are blank; the chain, free and target blocks carry exactly the printed observations and end at
`N`; exactly `K` recursive indices are unobserved; and `N ≤ c(n + 1)³` and every observed value
has at most `c(n + 1)³` bits. -/
theorem reduction_instance :
    ∃ c : ℕ, ∀ (G : NPGraph.Graph) (K : ℕ), NPGraph.IsSimple G → 1 ≤ numEdges G → 1 ≤ K →
      K + 1 ≤ G.order →
      (edgeList G).Nodup ∧
      (∀ e : ℕ × ℕ, e ∈ edgeList G ↔ e.1 < e.2 ∧ NPGraph.Adj G e.1 e.2) ∧
      (reductionInstance G K).IsValid ∧
      (reductionInstance G K).d = L0 G K ∧
      (reductionInstance G K).N = 2 * L0 G K - 1 ∧
      (reductionInstance G K).p = (reductionInstance G K).N + 1 ∧
      2 * (reductionInstance G K).d - 1 = (reductionInstance G K).N ∧
      (∀ i, i < (reductionInstance G K).d → (reductionInstance G K).obs.getD i none = none) ∧
      (∀ ℓ, 1 ≤ ℓ → ℓ ≤ zexp G.order →
        (reductionInstance G K).obs.getD ((reductionInstance G K).d + ℓ - 1) none =
          some (2 ^ (ℓ + 1) - 1)) ∧
      (∀ k, 1 ≤ k → k ≤ K →
        (reductionInstance G K).obs.getD ((reductionInstance G K).d + zexp G.order + k - 1) none =
          none) ∧
      (∀ k, 1 ≤ k → k ≤ numEdges G →
        (reductionInstance G K).obs.getD
            ((reductionInstance G K).d + zexp G.order + K + k - 1) none =
          some (2 * target G.order ((edgeList G).getD (k - 1) (0, 0)) - 1)) ∧
      (reductionInstance G K).d + zexp G.order + K + numEdges G - 1 = (reductionInstance G K).N ∧
      (reductionInstance G K).unobservedRecursive = K ∧
      (reductionInstance G K).N ≤ c * (G.order + 1) ^ 3 ∧
      (∀ i v, (reductionInstance G K).obs.getD i none = some v →
        Nat.size v ≤ c * (G.order + 1) ^ 3) := by sorry

end CallCountPOPBR
Source
Swain, Recursive Call-Count Sequences: Structural Realisability, Periodic Branching, and Partial-Observation Reconstruction, STOC 2027 submission (manuscript sha256 79135884dd262e4297079b8d2f7adb5deec1b329582779ed16decf34adae2411), Section 2.1, "The reduction.", p.5
Human review
  • Endorsed by Shuze Chen · Nov 6, 2026

    Confirmed by the moderator at approval.

  • Endorsed by iris · Nov 6, 2026

    Confirmed by the mission captain (proposal self-audit).

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