Lemma 7 — Pr[max_{i∈I^hi} T_i > b + 1 + α] ≤ 2^(2−α)/(α − 2) for α > 2
ProvedStochLoadBal.Makespan.lemma_7p2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1stochastic-schedulingtail-boundunion-bound
Under the hypotheses of Observation 3 (sizes measurable, integrable, jobs independent; feasible for the LP (5)–(9) with parameter ; an assignment with (13)–(14)), let be the set of machines of class at least and the truncated load of machine . Then for every ,
Integrating this tail bound gives , the first term of the bound in Lemma 6.
Formalization Note The event is written as "some machine of class at least has ", which is the event and is empty when .
Preamble
import Mathlib import Definitions.Def_StochLoadBal_Makespan_Rounding
Formal statement
namespace StochLoadBal.Makespan
open MeasureTheory ProbabilityTheory
/-- Lemma 7 (p. 13): for a feasible LP solution `y`, an assignment `σ` with (13)–(14), and
`α > 2`, the probability that some machine `i ∈ I^hi` (class `≥ 3`) has truncated load
`T_i = Σ_{j ∈ J_i} X'_{ij} > b + 1 + α` is at most `2^{2-α}/(α - 2)`. -/
theorem lemma_7 {m n : ℕ} (hm : 1 ≤ m) {Ω : Type} [MeasurableSpace Ω] (P : Measure Ω)
[IsProbabilityMeasure P] (X : Fin m → Fin n → Ω → ℝ)
(hmeas : ∀ i j, Measurable (X i j)) (hnn : ∀ i j ω, 0 ≤ X i j ω)
(hint : ∀ i j, Integrable (X i j) P) (hind : iIndepFun (fun j ω i => X i j ω) P)
(b : ℝ) (y : Fin m → Fin n → ℝ) (hy : LPFeasible P X b y)
(σ : Fin n → Fin m) (hσ : IsRoundingOutput P X b y σ) (α : ℝ) (hα : 2 < α) :
(P {ω | ∃ i, 3 ≤ machineClass P X b y i ∧
b + 1 + α < ∑ j ∈ Finset.univ.filter (fun j => σ j = i), trunc (X i j ω)}).toReal
≤ (2 : ℝ) ^ (2 - α) / (α - 2) := by sorry
end StochLoadBal.Makespan
Source
Gupta, Kumar, Nagarajan & Shen, Stochastic Load Balancing on Unrelated Machines, arXiv:1904.07271v1, p. 13, Lemma 7 (I^hi and T_i defined in the proof of Lemma 6, p. 12)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.