Skip to content

P-Matrix

A real square matrix whose every principal minor is strictly positive, equivalently guaranteeing a unique solution to every associated linear complementarity problem.

Version
v2 · 2026-09-06 · History
Domain-specific #
2443
Origin domain
mathematics
Subdomain
matrix theory
Aliases
P matrix, P-matrix class

Core Idea

A real square matrix \(A\in\mathbb R^{n\times n}\) is a P-matrix when every principal minor is strictly positive:

\[ \det A[I,I]>0\qquad\text{for every }I\subseteq\{1,\ldots,n\}. \]

This includes all diagonal entries and the full determinant. Classical work connects this hereditary positivity to generalized monotonicity and sign properties.[1] Most prominently, \(A\) is a P-matrix exactly when the linear complementarity problem

\[ w=Az+q\ge0,\quad z\ge0,\quad z^Tw=0 \]

has a unique solution for every \(q\in\mathbb R^n\).[2]

The recognition invariant is real square matrix + strict positivity of all principal minors + universal complementarity uniqueness.

Structural Signature

  • A real \(n\times n\) matrix.
  • Principal submatrices indexed by the same row and column set.
  • Strictly positive determinant for each nonempty principal submatrix.
  • Positive diagonal entries and positive determinant.
  • Nonsingularity of every principal submatrix.
  • Closure under simultaneous row/column permutation.
  • Closure under positive diagonal scaling and transpose.
  • Equivalent sign-nonreversal characterizations.
  • Unique LCP(A,q) solution for every right-hand side.
  • Spectral exclusion from a wedge around the negative real axis.
  • P₀-matrices as the nonnegative-principal-minor closure.

What It Is Not

A P-matrix need not be symmetric or positive definite. Symmetric P-matrices are positive definite, but nonsymmetric P-matrices can have complex eigenvalues and need not have all eigenvalues in the open right half-plane for arbitrary dimension.

It is not a matrix whose characteristic polynomial is positive, a polynomial in a matrix, or a stochastic transition matrix. A P₀-matrix permits zero principal minors and loses strict complementarity uniqueness.

Scope of Application

P-matrices characterize globally unique linear complementarity problems, arise in variational inequalities and equilibrium models, and supply sufficient Jacobian conditions for global injectivity on rectangular regions under Gale–Nikaidô-type theorems.[3] They also organize relationships among positive-definite, nonsingular M-, and sufficient matrices.

Testing every principal minor is exponential by direct enumeration, and recognizing P-matrices is computationally difficult in general; structured subclasses and alternative certificates matter in applications.

Clarity

State that the matrix is real and square, define principal rather than arbitrary minors, and distinguish strict P from P₀. If using an equivalent condition, cite its precise hypotheses. Do not infer positive stability solely from P-membership without additional structure.

Manages Complexity

One sign condition across nested coordinate restrictions guarantees well-posedness of an entire parametric family of complementarity problems. Equivalent characterizations let analysts switch among determinants, sign patterns, pivoting behavior, global univalence, and solution uniqueness according to the task.

Abstract Reasoning

  1. Verify real square shape.
  2. Choose a recognition route: principal minors, sign reversal, or complementarity.
  3. For small matrices, enumerate every principal index set.
  4. Use structure such as symmetry or Z-matrix signs when available.
  5. Separate P and P₀ boundary cases.
  6. Apply closure properties under transpose, permutation, or positive diagonal scaling.
  7. Invoke universal LCP uniqueness only for all \(q\), not one instance.
  8. Treat spectral conclusions with the exact wedge bound.
  9. Record numerical uncertainty near zero minors.

Knowledge Transfer

The portable pattern is local positivity on every coordinate-restricted subsystem guaranteeing global uniqueness of a constrained response problem. It transfers to hereditary certificates, monotonicity surrogates, injectivity tests, and equilibrium uniqueness. The proposed immediate parent is Matrix.

Examples

Positive-definite matrix. Every real symmetric positive-definite matrix has positive principal minors and is therefore a P-matrix.

Nonsymmetric example. \(\begin{pmatrix}1&2\\-1&1\end{pmatrix}\) has positive diagonal entries and determinant \(3\), so it is a P-matrix despite not being symmetric.

Complementarity. For a P-matrix \(A\), every vector \(q\) produces exactly one complementary pair \((z,w)\).[2]

Structural Tensions

  • Local principal minors versus global solution behavior.
  • Strict positivity versus boundary closure.
  • Symmetric intuition versus nonsymmetric spectra.
  • Exact determinant signs versus numerical conditioning.
  • Strong uniqueness theorem versus recognition complexity.
  • Matrix class inclusion versus non-equivalence.

Structural–Framed Character

Hereditary restriction, positivity, uniqueness, and invariance are structural. Principal minors, complementarity, Jacobians, and spectral wedges supply the constitutive matrix-theory frame.

Structural Core vs. Domain Accent

The portable core is a property inherited by every coordinate principal subsystem that certifies global uniqueness. The domain accent is strict determinant positivity for real matrices and its complementarity equivalences.

Matrix is the proposed immediate parent. Constraint, Positivity, Uniqueness, Substructure, Equilibrium, Determinant, and Stability are related. Kellogg's theorem bounds P-matrix eigenvalue arguments without asserting ordinary positive definiteness.[4]

The prospective queue contains one strict edge to domain_specific:matrix. No live DAG mutation is authorized.

Relationships to Other Abstractions

Local relationship map for P-MatrixParents appear above the current abstraction, mutual partners to the right, and children below. Node labels state whether each abstraction is prime or domain-specific; colors identify relation types.P-MatrixDOMAINDomain-specific abstraction: Matrix — is a kind ofMatrixDOMAIN

Current abstraction P-Matrix Domain-specific

Parents (1) — more general patterns this builds on

  • P-Matrix is a kind of Matrix Domain-specific

    Matrix is the proposed immediate parent.

Hierarchy paths (5) — routes to 5 parentless roots

Neighborhood in Abstraction Space

P-Matrix sits in a sparse region of the domain-specific corpus (90th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Unclustered & Miscellaneous (1565 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-09-08

Not to Be Confused With

  • Positive-definite matrix.
  • Positive matrix entrywise.
  • P₀-matrix.
  • M-matrix.
  • Polynomial matrix.
  • Permutation matrix.
  • Matrix similarity class.
  • Having positive determinant alone.

References

[1] Miroslav Fiedler and Vlastimil Pták, “Some Generalizations of Positive Definiteness and Monotonicity,” Numerische Mathematik 9 (1966): 163–172, doi:10.1007/BF02166034. registry ↩

[2] Richard W. Cottle, Jong-Shi Pang, and Richard E. Stone, The Linear Complementarity Problem (SIAM Classics, 2009; original 1992), doi:10.1137/1.9780898719000. registry ↩a ↩b

[3] David Gale and Hukukane Nikaidô, “The Jacobian Matrix and Global Univalence of Mappings,” Mathematische Annalen 159 (1965): 81–93, doi:10.1007/BF01360282. registry ↩

[4] R. B. Kellogg, “On Complex Eigenvalues of M and P Matrices,” Numerische Mathematik 19 (1972): 170–175, doi:10.1007/BF01402527. registry ↩