Skip to content

Schur decomposition

A unitary similarity that transforms any complex square matrix to upper triangular form, placing its eigenvalues on the diagonal while preserving numerical stability.

Version
v1 · 2026-09-28 · History
Domain-specific #
11905
Domain group
Formal Sciences
Origin domain
Mathematics
Subdomains
Linear Algebra, Matrix Decompositions, Numerical Linear Algebra → Mathematics

Core Idea

The Schur decomposition states that every complex square matrix A can be written A = Q U Q*, where Q is unitary and U is upper triangular. Similarity preserves the spectrum, so the diagonal entries of U are the eigenvalues of A, even when A lacks a full basis of eigenvectors.

The columns of Q provide an orthonormal basis adapted to a complete flag of invariant subspaces. Upper-triangular entries above the diagonal retain the coupling that prevents general matrices from being unitarily diagonalized. Eigenvalues can be reordered, and repeated eigenspaces permit further basis choices, so the decomposition is not generally unique.

For a real matrix, an orthogonal basis produces a real quasi-triangular form with one-by-one blocks for real eigenvalues and two-by-two blocks for conjugate pairs. Computational eigenvalue algorithms often seek this form because unitary transformations preserve norms and are numerically preferable to ill-conditioned eigenvector bases.

Structural Signature

Sig role-phrases:

  • square matrix or operator. Supplies the finite-dimensional linear transformation to be reduced. Constitutive carrier. If altered: Rectangular matrices require a different decomposition such as SVD or QR.
  • unitary basis change. Preserves inner products and conditioning while changing coordinates. Constitutive transformation. If altered: A general similarity can triangularize but lacks the unitary stability property.
  • triangular Schur form. Places eigenvalues on the diagonal and coupling information above it. Identity-bearing output. If altered: Diagonal form is not guaranteed for defective or nonnormal matrices.
  • invariant flag. Orders nested invariant subspaces represented by leading basis vectors. Structural interpretation. If altered: Without invariance, upper-triangular leading blocks would not represent operator-stable subspaces.
  • real block convention. Keeps real arithmetic by representing complex-conjugate eigenpairs in two-by-two blocks. Field-dependent variant. If altered: Forcing a real triangular diagonal would lose nonreal eigenpairs.

What It Is Not

  • Not eigendecomposition. Triangular form exists even when the matrix is defective and not diagonalizable.
  • Not QR factorization. QR is A=QR; Schur is a unitary similarity A=QUQ*.
  • Not SVD. Singular values and two-sided unitary factors solve a different problem.
  • Not unique. Eigenvalue ordering and bases within invariant subspaces may vary.

Scope of Application

The factorization applies to finite-dimensional square matrices and underlies stable eigenvalue and invariant-subspace computations.

  • Eigenvalue algorithms. Provides the standard terminal triangular form.
  • Matrix functions. Evaluates functions through triangular recurrences.
  • Control theory. Separates stable and unstable invariant subspaces.
  • Perturbation analysis. Studies spectral sensitivity in an orthonormal basis.
  • Real arithmetic. Uses quasi-triangular blocks for conjugate pairs.

Clarity

Schur form separates existence of eigenvalues from existence of a well-conditioned eigenvector basis. It shows that every complex matrix can be reduced stably to triangular form, while only special matrices—such as normal ones—admit unitary diagonalization.

Manages Complexity

A dense matrix hides spectrum, invariant subspaces, and nonnormal coupling. The decomposition moves those into diagonal eigenvalues, leading invariant blocks, and strictly upper-triangular interaction, retaining full information under a norm-preserving basis change.

Abstract Reasoning

  1. Verify that the operator is square and choose complex or real field conventions.
  2. Compute an orthogonal or unitary reduction, commonly through QR-based iteration.
  3. Check triangular or quasi-triangular structure and reconstruction A=QUQ*.
  4. Read eigenvalues from diagonal entries or two-by-two real blocks.
  5. Reorder blocks or solve triangular equations only with conditioning and nonuniqueness in view.

Knowledge Transfer

The decomposition transfers literally across numerical linear-algebra problems involving square operators. ‘Triangularizing a problem’ elsewhere is analogy unless a genuine unitary similarity and invariant flag exist; the portable parent patterns are basis change, normal form, and decomposition.

Examples

Canonical

A defective complex matrix cannot be diagonalized, yet a unitary Q transforms it to upper triangular U. The repeated eigenvalue appears on U's diagonal and the nonzero superdiagonal records the remaining coupling.

Mapped back: square matrix or operator → defective A; unitary basis change → Q; triangular Schur form → U; invariant flag → leading Schur vectors; real block convention → not needed over C.

Applied / In Practice

A real control matrix is reduced to real Schur form and its one-by-one and two-by-two blocks are reordered so stable eigenvalues precede unstable ones. The corresponding leading Schur vectors span a stable invariant subspace.

Mapped back: square matrix or operator → state matrix; unitary basis change → orthogonal Q; triangular Schur form → quasi-triangular T; invariant flag → stable leading subspace; real block convention → two-by-two conjugate blocks.

Structural Tensions

T1: triangular existence vs. diagonal simplicity. Schur form always exists, while diagonalization may fail or be ill-conditioned. Diagnostic: Does the task need eigenvalues and invariant subspaces or individual eigenvectors?

T2: nonuniqueness vs. computational ordering. Equivalent Schur forms can order eigenvalues differently and choose different bases. Diagnostic: Which ordering supports the downstream calculation?

T3: real arithmetic vs. complex simplicity. Real blocks preserve real computation but make conjugate eigenvalues less visually direct. Diagnostic: Is a real quasi-triangular form or complex triangular form preferable?

Structural–Framed Character

Schur decomposition is structural within linear algebra. Its existence, similarity, and invariant flags are formal; ordering conventions are chosen. It is non-evaluative and independent of institution. The construction transfers among square matrices but its vocabulary remains mathematical. Its character: a norm-preserving triangular normal form exposing spectrum and invariant subspaces without demanding diagonalizability.

Structural Core vs. Domain Accent

Skeletal core. Change to a well-conditioned basis that reveals a nested invariant structure and a simpler normal form.

Domain-bound accent. Square matrices, unitary similarity, triangular factors, eigenvalues, and real conjugate blocks define Schur form.

Why not prime. Decomposition and normal form travel; the named theorem is a particular linear-algebraic construction.

  • Decomposition. A matrix is expressed through structured factors with an exact reconstruction law.
  • Change of basis. Unitary similarity preserves the operator while changing coordinates.
  • No canonical parent edge is asserted in the current DAG.

Neighborhood in Abstraction Space

Schur decomposition sits in a moderately populated region (48th percentile for distinctiveness): it has near-neighbors but no dense thicket of look-alikes.

Family — Unclustered & Miscellaneous (2551 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-08

Not to Be Confused With

  • Eigendecomposition. Tell: Is U diagonal, and does a full eigenvector basis exist?
  • QR decomposition. Tell: Is the equation a factorization A=QR or a similarity A=QUQ*?
  • Singular value decomposition. Tell: Are singular values and two different basis factors being computed?
  • Hessenberg form. Tell: Is the matrix only nearly triangular or in final Schur form?

References

  • Frozen Wikipedia discovery revision: https://en.wikipedia.org/wiki/Schur_decomposition (revision 1339751973).
  • Preserved source candidate: https://books.google.com/books?id=e-JFDAAAQBAJ&pg=PA177
  • Preserved source candidate: https://math.mit.edu/~gs/linearalgebra/ila5/lafe_schur03.pdf
  • Preserved source candidate: https://nhigham.com/2022/05/11/what-is-a-schur-decomposition/
  • Preserved source candidate: https://www.netlib.org/lapack/lug/

The frozen Wikipedia revision is discovery provenance. The retained source set was reviewed for identity, formal or operational relation, and scope. The encyclopedia's structural synthesis is bounded to those claims; a thin authority surface is recorded as a nonblocking source-strengthening repair rather than concealed.