Shift operator
Template:Short description Template:About
In mathematics, and in particular functional analysis, the shift operator, also known as the translation operator, is an operator that takes a function Template:Math to its translation Template:Math.<ref>{{#invoke:Template wrapper|{{#if:|list|wrap}}|_template=cite web |_exclude=urlname, _debug, id |url = https://mathworld.wolfram.com/{{#if:ShiftOperator%7CShiftOperator.html}} |title = Shift Operator |author = Weisstein, Eric W. |website = MathWorld |access-date = |ref = Template:SfnRef }}</ref> In time series analysis, the shift operator is called the lag operator.
Shift operators are examples of linear operators, important for their simplicity and natural occurrence. The shift operator action on functions of a real variable plays an important role in harmonic analysis, for example, it appears in the definitions of almost periodic functions, positive-definite functions, derivatives, and convolution.<ref name=mar>Template:Cite book</ref> Shifts of sequences (functions of an integer variable) appear in diverse areas such as Hardy spaces, the theory of abelian varieties, and the theory of symbolic dynamics, for which the baker's map is an explicit representation. The notion of triangulated category is a categorified analogue of the shift operator.
DefinitionEdit
Functions of a real variableEdit
The shift operator Template:Mvar (where Template:Tmath) takes a function Template:Mvar on Template:Tmath to its translation Template:Mvar,
- <math>T^t f(x) = f_t(x) = f(x+t)~.</math>
A practical operational calculus representation of the linear operator Template:Mvar in terms of the plain derivative Template:Tmath was introduced by Lagrange,
Template:Equation box 1~, </math> |cellpadding= 6 |border |border colour = #0073CF |bgcolor=#F9FFF7}}
which may be interpreted operationally through its formal Taylor expansion in Template:Mvar; and whose action on the monomial Template:Mvar is evident by the binomial theorem, and hence on all series in Template:Mvar, and so all functions Template:Math as above.<ref>Jordan, Charles, (1939/1965). Calculus of Finite Differences, (AMS Chelsea Publishing), Template:Isbn .</ref> This, then, is a formal encoding of the Taylor expansion in Heaviside's calculus.
The operator thus provides the prototype<ref>M Hamermesh (1989), Group Theory and Its Application to Physical Problems (Dover Books on Physics), Hamermesh ISBM 978-0486661810, Ch 8-6, pp 294-5, online.</ref> for Lie's celebrated advective flow for Abelian groups,
- <math> \exp\left(t \beta(x) \frac{d}{dx}\right) f(x) = \exp\left(t \frac{d}{dh}\right) F(h) = F(h+t) = f\left(h^{-1}(h(x)+t)\right),</math>
where the canonical coordinates Template:Mvar (Abel functions) are defined such that
- <math>h'(x)\equiv \frac 1 {\beta(x)} ~, \qquad f(x)\equiv F(h(x)). </math>
For example, it easily follows that <math>\beta (x)=x</math> yields scaling,
- <math> \exp\left(t x \frac{d}{dx}\right) f(x) = f(e^t x) , </math>
hence <math> \exp\left(i\pi x \tfrac{d}{dx}\right) f(x) = f(-x)</math> (parity); likewise, <math>\beta (x)=x^2</math> yields<ref>p 75 of Georg Scheffers (1891): Sophus Lie, Vorlesungen Ueber Differentialgleichungen Mit Bekannten Infinitesimalen Transformationen, Teubner, Leipzig, 1891. Template:Isbn online </ref>
- <math> \exp\left(t x^2 \frac{d}{dx}\right) f(x) = f \left(\frac{x}{1-tx}\right),</math>
<math>\beta (x)= \tfrac{1}{x}</math> yields
- <math> \exp\left(\frac{t} {x} \frac{d}{dx}\right) f(x) = f \left(\sqrt{x^2+2t} \right) ,</math>
<math>\beta (x)=e^x</math> yields
- <math> \exp\left (t e^x \frac d {dx} \right ) f(x) = f\left (\ln \left (\frac{1}{e^{-x} - t} \right ) \right ) ,</math>
etc.
The initial condition of the flow and the group property completely determine the entire Lie flow, providing a solution to the translation functional equation<ref name="acz">Aczel, J (2006), Lectures on Functional Equations and Their Applications (Dover Books on Mathematics, 2006), Ch. 6, Template:Isbn .</ref>
- <math>f_t(f_\tau (x))=f_{t+\tau} (x) .</math>
SequencesEdit
{{#invoke:Labelled list hatnote|labelledList|Main article|Main articles|Main page|Main pages}} The left shift operator acts on one-sided infinite sequence of numbers by
- <math> S^*: (a_1, a_2, a_3, \ldots) \mapsto (a_2, a_3, a_4, \ldots)</math>
and on two-sided infinite sequences by
- <math> T: (a_k)_{k\,=\,-\infty}^\infty \mapsto (a_{k+1})_{k\,=\,-\infty}^\infty.</math>
The right shift operator acts on one-sided infinite sequence of numbers by
- <math> S: (a_1, a_2, a_3, \ldots) \mapsto (0, a_1, a_2, \ldots)</math>
and on two-sided infinite sequences by
- <math> T^{-1}:(a_k)_{k\,=\,-\infty}^\infty \mapsto (a_{k-1})_{k\,=\,-\infty}^\infty.</math>
The right and left shift operators acting on two-sided infinite sequences are called bilateral shifts.
Abelian groupsEdit
In general, as illustrated above, if Template:Mvar is a function on an abelian group Template:Mvar, and Template:Mvar is an element of Template:Mvar, the shift operator Template:Mvar maps Template:Math to<ref name="acz" /><ref>"A one-parameter continuous group is equivalent to a group of translations". M Hamermesh, ibid.</ref>
- <math> F_g(h) = F(h+g).</math>
Properties of the shift operatorEdit
The shift operator acting on real- or complex-valued functions or sequences is a linear operator which preserves most of the standard norms which appear in functional analysis. Therefore, it is usually a continuous operator with norm one.
Action on Hilbert spacesEdit
The shift operator acting on two-sided sequences is a unitary operator on Template:Tmath The shift operator acting on functions of a real variable is a unitary operator on Template:Tmath
In both cases, the (left) shift operator satisfies the following commutation relation with the Fourier transform: <math display="block"> \mathcal{F} T^t = M^t \mathcal{F}, </math> where Template:Mvar is the multiplication operator by Template:Math. Therefore, the spectrum of Template:Mvar is the unit circle.
The one-sided shift Template:Mvar acting on Template:Tmath is a proper isometry with range equal to all vectors which vanish in the first coordinate. The operator Template:Mvar is a compression of Template:Math, in the sense that <math display="block">T^{-1}y = Sx \text{ for each } x \in \ell^2(\N),</math> where Template:Mvar is the vector in Template:Tmath with Template:Math for Template:Math and Template:Math for Template:Math. This observation is at the heart of the construction of many unitary dilations of isometries.
The spectrum of Template:Mvar is the unit disk. The shift Template:Mvar is one example of a Fredholm operator; it has Fredholm index −1.
GeneralizationEdit
Jean Delsarte introduced the notion of generalized shift operator (also called generalized displacement operator); it was further developed by Boris Levitan.<ref name = mar/><ref>Template:SpringerEOM</ref><ref>Template:SpringerEOM</ref>
A family of operators Template:Tmath acting on a space Template:Math of functions from a set Template:Mvar to Template:Tmath is called a family of generalized shift operators if the following properties hold:
- Associativity: let <math>(R^y f)(x) = (L^x f)(y).</math> Then <math>L^x R^y = R^y L^x.</math>
- There exists Template:Mvar in Template:Mvar such that Template:Mvar is the identity operator.
In this case, the set Template:Mvar is called a hypergroup.
See alsoEdit
- Arithmetic shift
- Logical shift
- Clock and shift matrices
- Finite difference
- Translation operator (quantum mechanics)
NotesEdit
BibliographyEdit
- Template:Cite book
- Marvin Rosenblum and James Rovnyak, Hardy Classes and Operator Theory, (1985) Oxford University Press.