Finite component-completion solve and explicit sparse output
ProvedNearlyOptimalPPR.Computation.root_component_solve_execLet the represented root active set be nonempty and consist of stored positive-degree vertices, with and valid epoch marks. The root component-solve routine has a finite execution whose explicit output denotes
Here the inverse is the principal-domain solve extended by zero outside . Writing , charged work is bounded by the explicit enumeration, sparse assembly, and emission allowance plus one exact-solve charge for dimension and sparse size . The instruction count has the corresponding allowance plus one primitive call, and the machine remains ready to continue.
This implementation contract is used when the controller has already established component completion; it separately charges output emission.
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 root_component_solve_exec (m : SolveModel) {n : ℕ} (g : InputGraph n) (ω : ℕ → Bool)
(logical : RecursiveState n) (s : Fin n) (α ε : ℝ) (ell length : ℕ) (epoch : ℤ)
(st : MachineState) (hh : st.scalar.halted=false)
(pm : RecursiveMachineMemory g s α ε ell length logical st) (hDepth : logical.top ≤ ell)
(hα : 0<α ∧ α<1) (hNonempty : (logical.frames 0).active.Nonempty)
(hStored : (logical.frames 0).active ⊆ logical.stored)
(hDegree : ∀ v ∈ (logical.frames 0).active, 0<g.graph.degree v)
(hEpoch : st.scalar.core.ptrs 53=epoch)
(hbound : ∀ v : Fin n, st.scalar.integerMemory (vertexRecordAddress (16*(ell+3)) v.val 7) ≤ epoch) :
let size := volumeNat g (logical.frames 0).active+(logical.frames 0).active.card
let cost := (10*(logical.frames 0).active.card+5)+((24*(16*(ell+3))+100)*size+28)+
((4*(16*(ell+3))+21)*(logical.frames 0).active.card+6)+3
∃ steps out, MachineEval m g ω rootComponentSolveCode st steps out ∧ steps ≤ cost+1 ∧
outputVector out.output=potentialOutput g.graph α (solveOn g α (logical.frames 0).active (LocalPush.e s)) ∧
out.work ≤ st.work+(cost : ℕ)+solveCharge m (logical.frames 0).active.card size ∧ out.scalar.halted=false := by sorry
end NearlyOptimalPPR.ComputationSource
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.