Self-adjoint operator: Difference between revisions
imported>Hendra I. Nurdin mNo edit summary |
imported>Hendra I. Nurdin (continuing) |
||
Line 1: | Line 1: | ||
{{subpages}} | {{subpages}} | ||
In [[mathematics]], a '''self-adjoint operator''' is a [[denseness|densely]] defined [[linear operator]] mapping a complex Hilbert space onto itself and which is invariant under the unary operation of taking the [[adjoint (operator)|adjoint]]. That is, if | In [[mathematics]], a '''self-adjoint operator''' is a [[denseness|densely]] defined [[linear operator]] mapping a [[complex number|complex]] [[Hilbert space]] onto itself and which is invariant under the unary operation of taking the [[adjoint (operator theory)|adjoint]]. That is, if ''A'' is an operator with a domain <math>\scriptstyle H_0</math> which is a dense subspace of a complex Hilbert space ''H'' then it is self-adjoint if <math>\scriptstyle A=A^*</math>, where <math>\scriptstyle A^*</math> denotes the adjoint operator of ''A''. Note that the adjoint of any densely defined linear operator is always well-defined (in fact, the denseness of the domain of an operator is necessary for the existence of its adjoint) and two operators ''A'' and ''B'' are said to be equal if they have a common domain and their values coincide on that domain. | ||
On an infinite dimensional Hilbert space, a self-adjoint operator can be thought of as the analogy of a real symmetric matrix (i.e., a matrix which is its own transpose) or a Hermitian matrix in (i.e., a matrix which is its own Hermitian transpose) when these matrices are viewed as (bounded) linear operators on <math>\mathbb{R}^n</math> and <math>\mathbb{C}^n</math>, respectively. | On an infinite dimensional Hilbert space, a self-adjoint operator can be thought of as the analogy of a real symmetric [[matrix]] (i.e., a matrix which is its own transpose) or a [[Hermitian matrix]] in (i.e., a matrix which is its own Hermitian transpose) when these matrices are viewed as (bounded) linear operators on <math>\scriptstyle \mathbb{R}^n</math> and <math>\scriptstyle \mathbb{C}^n</math>, respectively. | ||
== | ==Special properties of a self-adjoint operator== | ||
The self-adjointness of an operator entails that it has some special properties. Some of these properties include: | |||
1. The [[eigenvalue|eigenvalues]] of a self-adjoint operator are real. As a special well-known case, all eigenvalues of a real symmetric matrices and a complex Hermitian matrices are real. | |||
2. By the von Neumann’s [[spectral theorem]], any self-adjoint operator ''X'' (not necessarily bounded) can be represented as | |||
:<math> | |||
X=\int_{-\infty}^{\infty} x E^X(dx), | |||
</math> | |||
where <math>\scriptstyle E^X</math> is the associated [[spectral measure]] of X (a spectral measure is a Hilbert space projection operator-valued [[measure]]) | |||
3. By [[Stone’s Theorem]], for any self-adjoint operator ''X'' the one parameter unitary [[group]] <math>\scriptstyle U=\{U_t\}_{t \in \mathbb{R}}</math> defined by <math>\scriptstyle U_t=\int_{-\infty}^{\infty} e^{-itx}\, E^X(dx)</math>, where <math>\scriptstyle E^X</math> is the spectral measure of ''X'', satisfies: | |||
:<math> | |||
\frac{dU_t}{dt} u=-iXU_t u =U_t(-iX)u, | |||
</math> | |||
for all ''u'' in the domain of ''X''. One says that the operator ''-iX'' is the [[generator of a semigroup|generator]] of the group ''U'' and writes: <math>\scriptstyle U_t=e^{-itX},\,\,t \in \mathbb{R} </math>. | |||
==Examples of self-adjoint operators== | |||
Consider the complex Hilbert space <math>\scriptstyle L^2(\mathbb{R};\mathbb{C})</math> of all complex-valued square integrable functions on <math>\scriptstyle \mathbb{R}</math> with the complex inner product <math>\scriptstyle \langle f,g\rangle=\int_{-\infty}^{\infty}f(x)\overline{g(x)}\,dx</math>, and the dense subspace <math>\scriptstyle C^{\infty}_0(\mathbb{R};\mathbb{C}) </math> of <math>\scriptstyle L^2(\mathbb{R};\mathbb{C})</math> of all infinitely differentiable functions complex-valued functions on <math>\scriptstyle \mathbb{R}</math> vanishing at <math>\scriptstyle \pm \infty</math>. Define the operators ''Q'', ''P'' on <math>\scriptstyle C^{\infty}_0(\mathbb{R};\mathbb{C}) </math> as: | |||
:<math> | |||
Q(f)(x)= xf(x) \quad \forall f \in C^{\infty}_0(\mathbb{R};\mathbb{C}) | |||
</math> | |||
and | |||
:<math> | |||
P(f)(x)=i \hbar \frac{d}{dx}f(x) \quad \forall f \in C^{\infty}_0(\mathbb{R};\mathbb{C}), | |||
</math> | |||
where <math>\scriptstyle \hbar</math> is the real valued [[Planck's constant]]. Then ''Q'' and ''P'' are self-adjoint operators satisfying the commutation relation <math>\scriptstyle [Q,P]=i\hbar I</math> on <math>\scriptstyle C^{\infty}_0(\mathbb{R};\mathbb{C})</math>, where ''I'' denotes the identity operator. In [[quantum mechanics]], the pair ''Q'' and ''P'' is known as the [[Schrödinger representation]] of canonical conjugate position and momentum operators ''q'' and ''p'' satisfying the [[canonical commutation relation]] (CCR) <math>\scriptstyle [q,p]=i\hbar</math> on the Hilbert space <math>\scriptstyle L^2(\mathbb{R};\mathbb{C})</math>. | |||
==Further reading== | |||
#K. Yosida, ''Functional Analysis'' (6 ed.), ser. Classics in Mathematics, Berlin, Heidelberg, New York: Springer-Verlag, 1980. | |||
#K. Parthasarathy, ''An Introduction to Quantum Stochastic Calculus'', ser. Monographs in Mathematics, Basel, Boston, Berlin: Birkhauser Verlag, 1992. |
Revision as of 20:09, 9 November 2007
In mathematics, a self-adjoint operator is a densely defined linear operator mapping a complex Hilbert space onto itself and which is invariant under the unary operation of taking the adjoint. That is, if A is an operator with a domain which is a dense subspace of a complex Hilbert space H then it is self-adjoint if , where denotes the adjoint operator of A. Note that the adjoint of any densely defined linear operator is always well-defined (in fact, the denseness of the domain of an operator is necessary for the existence of its adjoint) and two operators A and B are said to be equal if they have a common domain and their values coincide on that domain.
On an infinite dimensional Hilbert space, a self-adjoint operator can be thought of as the analogy of a real symmetric matrix (i.e., a matrix which is its own transpose) or a Hermitian matrix in (i.e., a matrix which is its own Hermitian transpose) when these matrices are viewed as (bounded) linear operators on and , respectively.
Special properties of a self-adjoint operator
The self-adjointness of an operator entails that it has some special properties. Some of these properties include:
1. The eigenvalues of a self-adjoint operator are real. As a special well-known case, all eigenvalues of a real symmetric matrices and a complex Hermitian matrices are real.
2. By the von Neumann’s spectral theorem, any self-adjoint operator X (not necessarily bounded) can be represented as
where is the associated spectral measure of X (a spectral measure is a Hilbert space projection operator-valued measure)
3. By Stone’s Theorem, for any self-adjoint operator X the one parameter unitary group defined by , where is the spectral measure of X, satisfies:
for all u in the domain of X. One says that the operator -iX is the generator of the group U and writes: .
Examples of self-adjoint operators
Consider the complex Hilbert space of all complex-valued square integrable functions on with the complex inner product , and the dense subspace of of all infinitely differentiable functions complex-valued functions on vanishing at . Define the operators Q, P on as:
and
where is the real valued Planck's constant. Then Q and P are self-adjoint operators satisfying the commutation relation on , where I denotes the identity operator. In quantum mechanics, the pair Q and P is known as the Schrödinger representation of canonical conjugate position and momentum operators q and p satisfying the canonical commutation relation (CCR) on the Hilbert space .
Further reading
- K. Yosida, Functional Analysis (6 ed.), ser. Classics in Mathematics, Berlin, Heidelberg, New York: Springer-Verlag, 1980.
- K. Parthasarathy, An Introduction to Quantum Stochastic Calculus, ser. Monographs in Mathematics, Basel, Boston, Berlin: Birkhauser Verlag, 1992.