Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Source running-time bound transfers to actual recursive-controller execution

Proved
NearlyOptimalPPR.Computation.controller_source_run_certified

by Shuze Chen · Nov 19, 2026 · Mathlib 0df444a (Lean v4.33.1)

Suppose 0<α≤1/20<\alpha\le1/20<α≤1/2, 0<ε≤10<\varepsilon\le10<ε≤1, and the seed has positive degree. Assume the canonical recursive source run has a finite terminal prefix and source work at most EEE. Let a prepared concrete root represent its initial state with heap mass at most one. Choose queue capacity NNN, rank rrr, and array length LLL with

8(ℓ+1)E≤N<2r,r<L,8(\ell+1)E\le N<2^r,\qquad r<L,8(ℓ+1)E≤N<2r,r<L,

and sufficient cache-potential credit. Then the actual recursive controller has a finite execution, producing a nonnegative ACL-accurate explicit output. Its charged work is at most

Winitial+P+FE+1,W_{\rm initial}+P+F E+1,Winitial​+P+FE+1,

where PPP is the heap-cleaning coefficient and FFF the explicit compiler overhead factor. The source termination and work hypotheses are discharged separately when this transfer theorem is applied to the uniform algorithm.

The result derives all reachable allocation bounds and instruction charges from source finite-prefix invariants.

Preamble
import Definitions.Def_NearlyOptimalPPR_Computation_compiler_runtime
open Finset Matrix
open scoped ENNReal
set_option autoImplicit false
noncomputable section
attribute [local instance] Classical.propDecidable
Formal statement
namespace NearlyOptimalPPR.Computation
open SparseQueue
theorem controller_source_run_certified (m : SolveModel) {n : ℕ} (g : InputGraph n) (ω : ℕ → Bool)
    (s : Fin n) (α ε E : ℝ) (ell length N rank Q : ℕ)
    (hα : 0<α ∧ α≤1/2) (hε : 0<ε ∧ ε≤1) (hs : 0<g.graph.degree s)
    (hBound : recursiveWork m g s α ε ell [] ≤ ENNReal.ofReal E)
    (hWide : ((8*(ell+1) : ℕ) : ℝ≥0∞)*ENNReal.ofReal E ≤ (N : ℝ≥0∞))
    (hTermination : ∃ t, (recursiveRun m g s α (rootThreshold ε ell) ell [] t).done=true)
    (hCredit : 4*(16*(ell+3))+17 ≤ Q) (hCap : rank<length) (hPow : N<2^rank)
    (st : MachineState) (memory : ControllerMemory g s α ε ell length
      (recursiveInit g s (rootThreshold ε ell) ell []) st) (hM : memory.mass ≤ 1) :
    let P := cleanBodyBound (16*(ell+3)) length rank+2
    let F := compilerOverheadFactor ell length rank Q
    ∃ steps out, MachineEval m g ω recursiveControllerCode st steps out ∧
      (∀ v : Fin n, 0 ≤ outputVector out.output v) ∧ ACL g.graph α ε s (outputVector out.output) ∧
      out.work ≤ st.work+(P : ℝ≥0∞)+(F : ℝ≥0∞)*ENNReal.ofReal E+1 := by sorry
end NearlyOptimalPPR.Computation
Source
Li, Xu, Yang, Zhou, Nearly optimal Personalized PageRank computation, STOC 2027 submission #884, https://github.com/sigact-stoc/stoc2027/releases/download/papers-1/884.pdf, Algorithms 1 and 6–9, pp. 5,14–22; charged local-machine realization of the mission computational model.

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