Linear complementarity has a unique solution for every input iff the matrix is a P-matrix
ProvedRobinsonSR.Schur.lcp_unique_iff_pmatrixlinear-complementarityp-matrixp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1principal-minors
For a real square matrix , the linear complementarity problem with input asks for such that and . It has exactly one solution for every if and only if every nonempty principal minor of is positive:
This is the complementarity characterization cited by Robinson for the nonnegative orthant case.
Preamble
import Mathlib import Definitions.Def_RobinsonSR_Schur_Setting
Formal statement
namespace RobinsonSR.Schur
/-- Robinson, p. 52, citing [6], for the LCP (3.8). -/
theorem lcp_unique_iff_pmatrix (s : ℕ) (M : Matrix (Fin s) (Fin s) ℝ) :
(∀ z : Fin s → ℝ, ∃! w : Fin s → ℝ,
(∀ j, 0 ≤ (M.mulVec w - z) j) ∧
(∀ j, 0 ≤ w j) ∧
w ⬝ᵥ (M.mulVec w - z) = 0) ↔ IsPMatrix M := by sorry
end RobinsonSR.Schur
Source
Robinson, Strongly regular generalized equations, Math. Oper. Res. 5 (1980), p. 52, (3.8) and citation [6]
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.