Proof of Theorem 1 — one-iteration contraction
ProvedAdaptiveStepIPM.PredCorr.contractioninterior-point-methodslinear-programmingp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1
For , one Algorithm 1 iteration starting from reduces the duality measure according to
This is the per-iteration estimate displayed in the proof of Theorem 1 and drives its termination count.
Formalization Note The bound on used in the printed proof fails at ; this milestone records the range . Nothing is lost: at the predictor term vanishes, the admissible predictor steps form with no greatest element, and Algorithm 1 has no iteration at all.
Preamble
import Mathlib import Definitions.Def_AdaptiveStepIPM_PredCorr_Algorithm
Formal statement
namespace AdaptiveStepIPM.PredCorr
/-- The contraction displayed in the proof of Theorem 1, printed p. 9. -/
theorem contraction {m n : ℕ} (hn : 2 ≤ n)
(A : Matrix (Fin m) (Fin n) ℝ) (b : Fin m → ℝ) (c : Fin n → ℝ)
(x s xNext sNext : Fin n → ℝ)
(hxs : N2 A b c (1 / 4) x s)
(hstep : Alg1Step A b c x s xNext sNext) :
mu xNext sNext ≤
(1 - (8 : ℝ) ^ (-(1 / 4 : ℝ)) * (n : ℝ) ^ (-(1 / 2 : ℝ))) * mu x s := by sorry
end AdaptiveStepIPM.PredCorr
Source
Mizuno, Todd, Ye, On adaptive-step primal-dual interior-point algorithms for linear programming, Cornell ORIE Technical Report No. 944 (1990, rev. 1991), p. 9, proof of Theorem 1, contraction display
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.