04.01.2015 Views

FRACTIONAL BROWNIAN VECTOR FIELDS 1. Introduction. A one ...

FRACTIONAL BROWNIAN VECTOR FIELDS 1. Introduction. A one ...

FRACTIONAL BROWNIAN VECTOR FIELDS 1. Introduction. A one ...

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

<strong>FRACTIONAL</strong> <strong>BROWNIAN</strong> <strong>VECTOR</strong> <strong>FIELDS</strong> 5<br />

standard Cartesian coordinates) of some related vector differential operators. 3<br />

grad<br />

div<br />

curl<br />

graddiv<br />

curlcurl<br />

∆<br />

E<br />

(−∆) γ ξ<br />

<br />

←→<br />

<br />

jω;<br />

←→ (jω) T ;<br />

<br />

←→<br />

<br />

0 −jω3 jω 2<br />

jω 3 0 −jω 1<br />

−jω 2 jω 1 0<br />

←→ −ωω T ;<br />

<br />

←→ ‖ω‖ 2 I − ωω T ;<br />

<br />

←→ −‖ω‖ 2 I;<br />

<br />

<br />

;<br />

←→ ωω T /‖ω‖ 2 ;<br />

<br />

←→ ˆΦ<br />

<br />

γ<br />

(ω) = ξ ‖ω‖2γ e ξ irr ωωT + e ξ ‖ω‖ 2 sol<br />

I − ωωT<br />

‖ω‖ 2 <br />

.<br />

(2.8)<br />

The penultimate operator (E) and its complement (Id − E) project a vector field respectively<br />

onto its curl-free and divergence-free comp<strong>one</strong>nts. In other words, together they afford a<br />

Helmholtz decomposition of the vector field on which they act (these operators appear<br />

prominently in fluid dynamics literature [8–10, 46]). This is because<br />

In addition, <strong>one</strong> has<br />

div (Id − E) = 0 and curl E = 0.<br />

E grad = grad and E curl = 0.<br />

(Id − E) is known as the Leray projector in turbulence literature.<br />

Our notation for the fractional vector Laplacian (−∆) γ is motivated by the observation<br />

ξ<br />

that it can be factorized as<br />

(−∆) γ ξ = (−∆)γ 0<br />

<br />

e<br />

ξ irrE + e<br />

ξ sol (Id − E)<br />

<br />

.<br />

In view of the properties of the operator E, this factorization means that the operator<br />

(−∆) γ combines a coordinate-wise fractional Laplacian with a re-weighting of the curl- and<br />

ξ<br />

divergence-free comp<strong>one</strong>nts of the operand.<br />

2.2. Some properties of ˆΦ γ . Let us now take a closer look at the family of matrixvalued<br />

functions ˆΦ γ , γ ∈ , ξ eξ irr ,e<br />

ξ<br />

ξ sol ∈ . They, of course, satisfy the required invariances:<br />

ˆΦ γ ξ (Ωω) = ΩˆΦ γ ξ (ω)ΩT ;<br />

ˆΦ γ ξ (σω) = σ 2γ ˆΦγ<br />

ξ (ω).<br />

But they exhibit, in addition, the following properties.<br />

(ˆΦ1) Closedness under multiplication: We have<br />

ˆΦ γ 1<br />

ξ 1<br />

(ω)ˆΦ γ 2<br />

ξ 2<br />

(ω) = ˆΦ γ 1 +γ 2<br />

ξ 1 +ξ 2<br />

(ω),<br />

3 Note that, while the curl operator is classically defined in three dimensions, the equivalents of graddiv,<br />

curlcurl, and ∆ = graddiv−curlcurl can be defined in any number of dimensions, for instance by their<br />

Fourier symbols. More precisely, for arbitrary d, the equivalents of −curlcurl and −graddiv that appear<br />

in the definition of the vector Laplacian correspond, respectively, to the product of d-dimensional curl and<br />

divergence with their adjoints.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!