where A^{*} is the conjugate transpose of A (if A is a real matrix, this is the same as the transpose of A).
Examples of normal matrices are unitary matrices, hermitian matrices and positive definite matrices.
It is useful to think of normal matrices in analogy to complex numbers, invertible normal matrices in analogy to nonzero complex numbers, the conjugate transpose in analogy to the complex conjugate, unitary matrices in analogy to complex numbers of absolute value 1, hermitian matrices in analogy to real numbers and positive definite matrices in analogy to positive real numbers.
The concept of normality is mainly important because normal matrices are precisely the ones to which the spectral theorem applies; in other words, normal matrices are precisely those matrices that can be represented by a diagonal matrix with respect to a properly chosen orthonormal basis of C^{n}. Phrased differently: a matrix is normal if and only if its eigenspaces span C^{n} and are pairwise orthogonal with respect to the standard inner product of C^{n}.
In general, the sum or product of two normal matrices need not be normal. However, if A and B are normal with AB = BA, then both AB and A + B are also normal and furthermore we can simultaneously diagonalize A and B in the following sense: there exists a unitary matrix U such UAU^{*} and UBU^{*} are both diagonal matrices. In this case, the columns of U^{*} are eigenvectors of both A and B and form an orthonormal basis of C^{n}.
If A is an invertible normal matrix, then there exists a unitary matrix U and a positive definite matrix R such that A = RU = UR. The matrices R and U are uniquely determined by A. This statement can be seen as an analog (and generalization) of the polar representation of nonzero complex numbers.
The concept of normal matrices can be generalized to normal operators on Hilbert spaces and to normal elements in Cstar algebras.
Search Encyclopedia

Featured Article
