Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 1 (Rounding Theorem) — every vertex of (LP) rounds to a 0-1 solution of (IP) on its support

Proved
UnrelatedSched.TwoApprox.theorem_1

by mikedeng1 · Oct 6, 2026 · Mathlib 0df444a (Lean v4.33.1)

linear-programmingp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1roundingscheduling

Let P=(pij)∈Nm×nP=(p_{ij})\in\mathbb N^{m\times n}P=(pij​)∈Nm×n, deadlines d1,…,dm∈Rd_1,\dots,d_m\in\mathbb Rd1​,…,dm​∈R and a threshold t≥0t\ge 0t≥0. Suppose the linear program

∑i∈Mj(t)xij=1 (j=1,…,n),∑j∈Ji(t)pijxij≤di (i=1,…,m),xij≥0 (j∈Ji(t))(LP)\sum_{i\in M_j(t)}x_{ij}=1\ (j=1,\dots,n),\qquad\sum_{j\in J_i(t)}p_{ij}x_{ij}\le d_i\ (i=1,\dots,m),\qquad x_{ij}\ge 0\ (j\in J_i(t))\qquad\text{(LP)}i∈Mj​(t)∑​xij​=1 (j=1,…,n),j∈Ji​(t)∑​pij​xij​≤di​ (i=1,…,m),xij​≥0 (j∈Ji​(t))(LP)

is feasible, and let x~\tilde xx~ be any vertex of this polytope. Then x~\tilde xx~ can be rounded to a feasible solution xˉ\bar xxˉ of the integer program

∑i∈Mj(t)xij=1,∑j∈Ji(t)pijxij≤di+t,xij∈{0,1}(IP)\sum_{i\in M_j(t)}x_{ij}=1,\qquad\sum_{j\in J_i(t)}p_{ij}x_{ij}\le d_i+t,\qquad x_{ij}\in\{0,1\}\qquad\text{(IP)}i∈Mj​(t)∑​xij​=1,j∈Ji​(t)∑​pij​xij​≤di​+t,xij​∈{0,1}(IP)

with xˉij=1\bar x_{ij}=1xˉij​=1 only where x~ij>0\tilde x_{ij}>0x~ij​>0. In schedule form: there is a schedule σ\sigmaσ with x~σ(j)j>0\tilde x_{\sigma(j)j}>0x~σ(j)j​>0 for every job, pσ(j)j≤tp_{\sigma(j)j}\le tpσ(j)j​≤t, and load at most di+td_i+tdi​+t on every machine iii.

The Rounding Theorem is the key tool of the paper: it converts a fractional solution of the deadline LP into a schedule that overshoots each deadline by at most ttt, the largest processing time allowed.

Formalization Note The paper states "and this rounding can be done in polynomial time"; running time is not formalized. "Rounding" is formalized as support containment (xˉij=1⇒x~ij>0\bar x_{ij}=1\Rightarrow\tilde x_{ij}>0xˉij​=1⇒x~ij​>0), which in particular keeps every integral entry x~ij=1\tilde x_{ij}=1x~ij​=1. The paper takes d∈Z+md\in\mathbb Z_+^md∈Z+m​ and t∈Z+t\in\mathbb Z_+t∈Z+​; the proof never uses integrality, so real ddd and real t≥0t\ge 0t≥0 are allowed. Feasibility of (LP) is implied by the existence of a vertex.

Preamble
import Mathlib
import Definitions.Def_MatousekLP_Scheduling_Schedule
import Definitions.Def_UnrelatedSched_TwoApprox_DeadlineLP
Formal statement
namespace UnrelatedSched.TwoApprox

/-- Theorem 1 (Rounding Theorem), p. 3: every vertex `x̃` of (LP) can be rounded to a 0-1 solution of
(IP) supported on `x̃`. -/
theorem theorem_1 {m n : ℕ} (P : Matrix (Fin m) (Fin n) ℕ) (d : Fin m → ℝ) (t : ℝ) (ht : 0 ≤ t)
    (x : Matrix (Fin m) (Fin n) ℝ) (hx : IsLPVertex P d t x) :
    ∃ σ : Fin n → Fin m, SupportedOn x σ ∧ IPFeasible P d t σ := by sorry

end UnrelatedSched.TwoApprox
Source
Lenstra, Shmoys, Tardos, Approximation algorithms for scheduling unrelated parallel machines, CWI Report OS-R8714 (1987), p. 3, Theorem 1
Human review
  • Endorsed by Shuze Chen · Oct 6, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 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