Ordered exponential

From testwiki
Revision as of 02:45, 27 May 2024 by imported>Cewbot (Fixing broken anchor: Reminder of an inactive anchor: geometric integral)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Template:More citations needed The ordered exponential, also called the path-ordered exponential, is a mathematical operation defined in non-commutative algebras, equivalent to the exponential of the integral in the commutative algebras. In practice the ordered exponential is used in matrix and operator algebras. It is a kind of product integral, or Volterra integral.

Definition

Let Template:Math be an algebra over a field Template:Math, and Template:Math be an element of Template:Math parameterized by the real numbers,

a:A.

The parameter Template:Mvar in Template:Math is often referred to as the time parameter in this context.

The ordered exponential of Template:Math is denoted

OE[a](t)𝒯{e0ta(t)dt}n=01n!0tdt'10tdt'n𝒯{a(t'1)a(t'n)}=n=00tdt'10t'1dt'20t'2dt'30t'n1dt'na(t'n)a(t'1)

where the term Template:Math is equal to 1 and where 𝒯 is the time-ordering operator. It is a higher-order operation that ensures the exponential is time-ordered, so that any product of Template:Math that occurs in the expansion of the exponential is ordered such that the value of Template:Mvar is increasing from right to left of the product. For example:

𝒯{a(1.2)a(9.5)a(4.1)}=a(9.5)a(4.1)a(1.2).

Time ordering is required, as products in the algebra are not necessarily commutative.

The operation maps a parameterized element onto another parameterized element, or symbolically,

OE:(A)(A).

There are various ways to define this integral more rigorously.

Product of exponentials

The ordered exponential can be defined as the left product integral of the infinitesimal exponentials, or equivalently, as an ordered product of exponentials in the limit as the number of terms grows to infinity:

OE[a](t)=0tea(t)dtlimN(ea(tN)Δtea(tN1)Δtea(t1)Δtea(t0)Δt)

where the time moments Template:Math are defined as Template:Math for Template:Math, and Template:Math.

The ordered exponential is in fact a geometric integralTemplate:Broken anchor.[1][2][3]

Solution to a differential equation

The ordered exponential is unique solution of the initial value problem:

ddtOE[a](t)=a(t)OE[a](t),OE[a](0)=1.

Solution to an integral equation

The ordered exponential is the solution to the integral equation:

OE[a](t)=1+0ta(t)OE[a](t)dt.

This equation is equivalent to the previous initial value problem.

Infinite series expansion

The ordered exponential can be defined as an infinite sum,

OE[a](t)=1+0ta(t1)dt1+0tdt10t1dt2a(t1)a(t2)+.

This can be derived by recursively substituting the integral equation into itself.

Example

Given a manifold M where for a eTM with group transformation g:ege it holds at a point xM:

de(x)+J(x)e(x)=0.

Here, d denotes exterior differentiation and J(x) is the connection operator (1-form field) acting on e(x). When integrating above equation it holds (now, J(x) is the connection operator expressed in a coordinate basis)

e(y)=Pexp(xyJ(γ(t))γ(t)dt)e(x)

with the path-ordering operator P that orders factors in order of the path γ(t)M. For the special case that J(x) is an antisymmetric operator and γ is an infinitesimal rectangle with edge lengths |u|,|v| and corners at points x,x+u,x+u+v,x+v, above expression simplifies as follows :

OE[J]e(x)=exp[J(x+v)(v)]exp[J(x+u+v)(u)]exp[J(x+u)v]exp[J(x)u]e(x)=[1J(x+v)(v)][1J(x+u+v)(u)][1J(x+u)v][1J(x)u]e(x).

Hence, it holds the group transformation identity OE[J]gOE[J]g1. If J(x) is a smooth connection, expanding above quantity to second order in infinitesimal quantities |u|,|v| one obtains for the ordered exponential the identity with a correction term that is proportional to the curvature tensor.

See also

References

  1. Michael Grossman and Robert Katz. Non-Newtonian Calculus, Template:ISBN, 1972.
  2. A. E. Bashirov, E. M. Kurpınar, A. Özyapıcı. Multiplicative calculus and its applications, Journal of Mathematical Analysis and Applications, 2008.
  3. Luc Florack and Hans van Assen."Multiplicative calculus in biomedical image analysis", Journal of Mathematical Imaging and Vision, 2011.