(62), Appendix p. 38 — if w₁² − ‖w_{2:n}‖² = 1, the arrow form of (31) squares to −Q + 2wwᵀ
ProvedConicQuadIPM.NTScaling.eq_62conic-optimizationnesterov-todd-scalingp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1second-order-cone
Let , let , and let satisfy and . With
one has
This is the identity (33) for the quadratic cone and the step that lets the NT condition be solved for .
Formalization Note Vectors are Fin d → ℝ with 0-based indices (the paper's is coord x 0); is tailSq x 1, a sum of squares (Euclidean, never Mathlib's sup norm); is x ⬝ᵥ s. The hypothesis makes the division defined; under the normalization it only excludes .
Preamble
import Mathlib import Definitions.Def_ConicQuadIPM_NTScaling_Setting import Definitions.Def_ConicQuadIPM_NTScaling_Scaling open Matrix
Formal statement
namespace ConicQuadIPM.NTScaling
/-- (62), Appendix, proof of Lemma 4.2, p. 38: if `w₁² − ‖w_{2:n}‖² = 1`, the arrow-form matrix
of (31) squares to `−Q + 2wwᵀ` (quadratic cone). -/
theorem eq_62 (d : ℕ) (hd : 1 ≤ d) (w : Fin d → ℝ)
(hnorm : ConicQuadIPM.Complementarity.coord w 0 ^ 2 - ConicQuadIPM.Complementarity.tailSq w 1 = 1) (hw : 0 < 1 + ConicQuadIPM.Complementarity.coord w 0) :
WquadArrow w * WquadArrow w = -Qmat .quad d + (2 : ℝ) • vecMulVec w w := by sorry
end ConicQuadIPM.NTScaling
Source
Andersen, Roos & Terlaky, On implementing a primal-dual interior-point method for conic quadratic optimization, Math. Program. (2003), DOI 10.1007/s10107-002-0349-3; authors' preprint of 18 Dec 2000, p. 38, Appendix, proof of Lemma 4.2, (62)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.