Hermitian matrix

Hermitian matrix
Name	Hermitian matrix
Field	Linear algebra

Hermitian matrix is a square matrix equal to its own conjugate transpose, central in linear algebra and functional analysis. It arises in quantum mechanics, signal processing, and numerical analysis and connects to operators studied by David Hilbert, John von Neumann, Évariste Galois, Carl Friedrich Gauss, and institutions such as Princeton University and University of Göttingen in the development of spectral theory. Hermitian matrices underpin results proven by contributors like Issai Schur, Friedrich Wilhelm Joseph von Struve, and Stefan Banach and are implemented in software from organizations including National Institute of Standards and Technology and IBM.

Definition and basic properties

A Hermitian matrix is defined over the complex numbers as a square matrix A satisfying A = A*, where A* denotes the conjugate transpose; this definition was formalized in work associated with David Hilbert and John von Neumann. Fundamental properties include real diagonal entries, self-adjointness in the matrix algebra context explored at University of Cambridge and Harvard University, and bilinear form relations connected to the studies of Bernhard Riemann and Sofia Kovalevskaya. Hermitian matrices are closed under linear combinations with real coefficients, relate to positive semidefinite matrices investigated by Marcel Riesz and Helmut Hasse, and appear in canonical problems studied at École Normale Supérieure and Massachusetts Institute of Technology.

Spectral theorem and eigenvalues

The spectral theorem for Hermitian matrices guarantees a basis of orthonormal eigenvectors and real eigenvalues, a central result in work by John von Neumann, Erwin Schrödinger, and Paul Dirac. Applications of the spectral theorem connect to expansions used in research at CERN and in formulations by Niels Bohr and Werner Heisenberg. Eigenvalue interlacing and variational characterizations trace to contributions by Rayleigh and Ritz and are exploited in algorithms developed at Bell Labs and Los Alamos National Laboratory. The distribution and multiplicity of eigenvalues are key in stability analyses found in publications from Stanford University and California Institute of Technology.

Operations and algebraic properties

Algebraic operations preserve Hermitian structure under key conditions: sums of Hermitian matrices and real scalar multiples remain Hermitian, whereas products are Hermitian only when matrices commute, a property investigated in contexts involving Élie Cartan and Hermann Weyl. The commutator relations and matrix exponentials relate to representations used in studies by Sophus Lie and Édouard Goursat. Polar decompositions and singular value decompositions connect Hermitian factors to unitary matrices examined by Felix Klein and Hermann Grassmann. Determinant and trace properties tie to invariants considered by Camille Jordan and James Joseph Sylvester in invariant theory at University of Oxford.

Forms and examples

Canonical examples include real symmetric matrices studied by Augustin-Louis Cauchy and complex Hermitian matrices appearing in the Hamiltonians of Erwin Schrödinger and Paul Dirac. Diagonal matrices with real entries, Toeplitz Hermitian matrices investigated at ETH Zurich, and circulant Hermitian matrices analyzed in work from Niels Fabian Helge von Koch and Harald Bohr illustrate common forms. Block Hermitian matrices arise in control theory research at Imperial College London and INRIA. Special classes like positive definite Hermitian matrices connect to optimization problems considered by John Nash and Leonid Kantorovich.

Applications in physics and engineering

Hermitian matrices model observables in quantum mechanics per formulations by Paul Dirac, John von Neumann, and Erwin Schrödinger; they ensure measurable quantities are real and spectra correspond to physical outcomes investigated at Los Alamos National Laboratory and CERN. In electrical engineering and signal processing, Hermitian covariance matrices are central to algorithms developed at Bell Labs and MIT Lincoln Laboratory. Structural vibration analysis, modal decomposition, and stability studies at General Electric and Siemens exploit Hermitian properties. In control theory and optimization used by Boeing and Airbus, Riccati equations and Lyapunov methods frequently involve Hermitian matrices.

Generalizations and related concepts

Generalizations include self-adjoint operators on infinite-dimensional Hilbert spaces central to work by David Hilbert and Stefan Banach, normal matrices studied by Camille Jordan, and skew-Hermitian matrices related to Lie algebra elements in research by Sophus Lie and Élie Cartan. Connections to positive semidefinite kernels used in machine learning at Google and DeepMind extend Hermitian concepts to reproducing kernel Hilbert spaces developed by Nikolai Nikolaevich Luzin and James Mercer. Matrix pencils, generalized eigenvalue problems, and non-Hermitian perturbation theory are active research topics at Princeton University and University of California, Berkeley.

Category:Linear algebra