Cholesky factor ,
DefinitionMatrixLeeSidford_Supplement_cholFactorinterior-point-methodsp2o-gran-per-chapterp2o-plan-bookp2o-v1self-concordant-barrierssemidefinite-programming
For a positive definite matrix , its Cholesky factor is the lower triangular matrix with positive diagonal entries such that
It enters the stationarity matrix and the gradient matrix of Definition 8.1 and the damped inverse update (8.2) of the weight.
Formalization Note The factor is selected by choice among the lower triangular matrices with positive diagonal and ; it exists (and is unique) when . On an input with no such factor the value is the junk matrix ; every use in this mission is at a positive definite .
Definition code
import Mathlib
open Matrix
namespace MatrixLeeSidford.Supplement
/-- The **Cholesky factor** `L = chol(G)` of a positive definite `G`: the lower triangular matrix
with positive diagonal and `G = L Lᵀ` (Song–Zhang, *A Matrix Lee–Sidford Barrier and Faster Semidefinite Programming*, STOC 2027 submission manuscript of October 31, 2026 (sha256 7c638b45…412c1), Definition 8.1, p.53, "For G ≻ 0 with Cholesky
factor L"; Algorithm 8, line 2, p.82, "L = chol(G)").
**Formalization Note** Defined by choice: if some `L` is lower triangular (`L i j = 0` for
`i < j`), has positive diagonal and satisfies `L * Lᵀ = G`, `cholFactor G` is one such `L`
(it is unique, a theorem not used here); otherwise it is the junk value `0`. Every use in this
mission is at a positive definite `G`, where the factor exists. -/
noncomputable def cholFactor {m : ℕ} (G : Matrix (Fin m) (Fin m) ℝ) : Matrix (Fin m) (Fin m) ℝ :=
open Classical in
if h : ∃ L : Matrix (Fin m) (Fin m) ℝ,
(∀ i j, i < j → L i j = 0) ∧ (∀ i, 0 < L i i) ∧ L * Lᵀ = G then
h.choose
else 0
end MatrixLeeSidford.Supplement
Source
Song–Zhang, A Matrix Lee–Sidford Barrier and Faster Semidefinite Programming, STOC 2027 submission (manuscript of October 31, 2026, sha256 7c638b45…412c1), Definition 8.1, p.53; Algorithm 8, line 2, p.82