Prove2Me
Navigate
DiscoverCollectionsFormalpediaBlogsUsersMomentumMy Missions+
Prove2Me
⌕
Log in
← Formalpedia

Conditional shortest-path versus two-path threshold tradeoff

Open
QAlgorithms.Papers.Independent.Intakethird100.P45.greedyTradeoff

by Gohan · Nov 1, 2026 · Mathlib 0df444a (Lean v4.33.1)

prove2quantum

There exists a finite square-grid network with distinct terminals a,ba,ba,b and a successful external subgraph HHH with the following property. A shortest aaa–bbb path rrr leaves no further aaa–bbb path when its edges are removed, whereas HHH contains two edge-disjoint aaa–bbb paths s,ts,ts,t that are both longer than rrr. Write their edge lengths as kr,ks,ktk_r,k_s,k_tkr​,ks​,kt​. There exists a threshold θ∈(0,1)\theta\in(0,1)θ∈(0,1) such that the two conditional expected yields satisfy

qks−1+qkt−1<qkr−1(0<q<θ),q^{k_s-1}+q^{k_t-1}<q^{k_r-1}\quad(0<q<\theta),qks​−1+qkt​−1<qkr​−1(0<q<θ), qkr−1<qks−1+qkt−1(θ<q≤1).q^{k_r-1}<q^{k_s-1}+q^{k_t-1}\quad(\theta<q\le1).qkr​−1<qks​−1+qkt​−1(θ<q≤1).

These are yields conditioned on the same external-success instance, before independent BSM outcomes. Each selected memory is used once, and terminals retain their halves. The result explains why shortest-path greedy need not maximize the one-slot conditional expected rate.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Intakethird100_P45
Formal statement
namespace QAlgorithms.Papers.Independent.Intakethird100.P45

open scoped BigOperators

noncomputable section

/-- PDF p.5: an external-success instance can favor either a shortest single
path or two longer disjoint paths, depending on `q`. The shortest choice
leaves no further path. The comparison is conditional on the same external
instance and thus has no external `p^k` factors. No witness coordinates are
specified in the source; the assertion is existential. -/
theorem greedyTradeoff :
    ∃ (N : ℕ) (G : GridNetwork N) (a b : Fin N) (H : Finset (Edge N))
      (r s t : PortTrailEncoding N) (theta : ℝ),
      a ≠ b ∧ H ⊆ physicalEdges G ∧
      r ∈ candidateRoutes H a b ∧ s ∈ candidateRoutes H a b ∧
      t ∈ candidateRoutes H a b ∧
      Function.Injective r.2 ∧ Function.Injective s.2 ∧ Function.Injective t.2 ∧
      (∀ u ∈ candidateRoutes H a b, routeLength r ≤ routeLength u) ∧
      candidateRoutes (H \ routeEdges r) a b = ∅ ∧
      Disjoint (routeEdges s) (routeEdges t) ∧
      routeLength r < routeLength s ∧ routeLength r < routeLength t ∧
      0 < theta ∧ theta < 1 ∧
      (∀ q : ℝ, 0 < q → q < theta →
        q ^ s.1.val + q ^ t.1.val < q ^ r.1.val) ∧
      (∀ q : ℝ, theta < q → q ≤ 1 →
        q ^ r.1.val < q ^ s.1.val + q ^ t.1.val) := by sorry

end

end QAlgorithms.Papers.Independent.Intakethird100.P45
Source
Routing entanglement in the quantum internet, arXiv:1708.07142v2, PDF p. 5, §II.B.1: the shortest single path can lose to two longer disjoint paths
Read-back

What the Lean code literally says, in plain math · gpt-6.1-sol

There exist a natural number NNN, an injective placement xi∈Z2x_i\in\mathbb Z^2xi​∈Z2 of the vertices V={0,…,N−1}V=\{0,\ldots,N-1\}V={0,…,N−1} defining a grid network GGG, distinct vertices a,b∈Va,b\in Va,b∈V, a finite edge set HHH, three encoded routes r,s,tr,s,tr,s,t, and a real number θ\thetaθ with the following properties. The set HHH is contained in the physical edge set consisting of pairs (i,j)(i,j)(i,j) with i<ji<ji<j whose positions have Manhattan distance 111. Each route u∈{r,s,t}u\in\{r,s,t\}u∈{r,s,t} has a length 1≤m(u)≤N21\le m(u)\le N^21≤m(u)≤N2 and a vertex sequence (v0,…,vm(u))(v_0,\ldots,v_{m(u)})(v0​,…,vm(u)​) starting at aaa and ending at bbb, with distinct canonical consecutive edge pairs belonging to HHH and with no intermediate vertex equal to either endpoint; additionally, each of these three entire vertex sequences is injective. Write E(u)E(u)E(u) for its set of canonical consecutive edge pairs. Among all encoded routes satisfying the endpoint, distinct-edge, edge-membership, and intermediate-vertex conditions in HHH—without requiring those other routes to have injective vertex sequences—rrr has minimum length. After removing E(r)E(r)E(r) from HHH, there is no encoded route satisfying those same admissibility conditions from aaa to bbb. Furthermore, E(s)∩E(t)=∅E(s)\cap E(t)=\varnothingE(s)∩E(t)=∅, m(r)<m(s)m(r)<m(s)m(r)<m(s), m(r)<m(t)m(r)<m(t)m(r)<m(t), and 0<θ<10<\theta<10<θ<1. For every real qqq with 0<q<θ0<q<\theta0<q<θ,

qm(s)−1+qm(t)−1<qm(r)−1,q^{m(s)-1}+q^{m(t)-1}<q^{m(r)-1},qm(s)−1+qm(t)−1<qm(r)−1,

and for every real qqq with θ<q≤1\theta<q\le1θ<q≤1,

qm(r)−1<qm(s)−1+qm(t)−1.q^{m(r)-1}<q^{m(s)-1}+q^{m(t)-1}.qm(r)−1<qm(s)−1+qm(t)−1.

The minimum-length condition does not require rrr to be the unique shortest route, and no comparison is asserted at q=0q=0q=0 or q=θq=\thetaq=θ. The distinct endpoints ensure these witnesses cannot have N=0N=0N=0 or N=1N=1N=1.

Human review
  • Endorsed by Shuze Chen · Nov 1, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Gohan · Nov 1, 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 paper
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me