adjoint endomorphism
Definition (the bilinear case).
Let be afinite-dimensional vector space over a field , and a symmetric
, non-degenerate bilinear mapping, for examplea real inner product
. For an endomorphism
wedefine the adjoint
of relative to to be the endomorphism, characterized by
It is convenient to identify with a linear isomorphism in the sense that
We then have
To put it another way, gives anisomorphism between andthe dual , and theadjoint is the endomorphism of that corresponds to thedual homomorphism (http://planetmath.org/DualHomomorphism). Here is a commutative diagram
toillustrate this idea:
Relation to the matrix transpose.
Let be a basis of , and let be the matrix of relative to this basis, i.e.
Let denote the matrix of the inner productrelative to the same basis, i.e.
Then, the representing matrix of relative to the same basisis given by Specializing further, suppose that thebasis in question is orthonormal, i.e. that
Then, the matrix of issimply the transpose .
The Hermitian (sesqui-linear) case.
If is an endomorphism of a unitary space (a complexvector space equipped with a Hermitian inner product (http://planetmath.org/HermitianForm)). In this setting we can define we definethe Hermitian adjoint by means of the familiaradjointness condition
However, the analogous operation at the matrix level is the conjugatetranspose
. Thus, if is the matrix of relative to an orthonormal basis
, then is thematrix of relative to the same basis.