14.11.2012 Views

Conformal Geometric Algebra in Stochastic Optimization Problems ...

Conformal Geometric Algebra in Stochastic Optimization Problems ...

Conformal Geometric Algebra in Stochastic Optimization Problems ...

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.

3.3. CONFORMAL ANALYTIC GEOMETRY 99<br />

3.3.7 Planes<br />

Planes are the last vector valued elements that are mentioned. Recall that a plane<br />

can be considered a special sphere with <strong>in</strong>f<strong>in</strong>ite radius. This notion has a solid<br />

base; replac<strong>in</strong>g the fourth po<strong>in</strong>t d <strong>in</strong> equation (3.21), the OPNS representation of<br />

a sphere, with ‘the po<strong>in</strong>t’ e, should consistently yield a sphere pass<strong>in</strong>g through<br />

<strong>in</strong>f<strong>in</strong>ity. The result<strong>in</strong>g OPNS expression does <strong>in</strong> effect match the plane equation<br />

(3.20).<br />

The IPNS of planes representation is succ<strong>in</strong>ct<br />

P = ˆn + de, d ≥ 0 (3.29)<br />

where ˆn, ||ˆn|| = 1, denotes the normal vector of the plane and d stands for the<br />

distance to the orig<strong>in</strong> eo. If the plane is not normalized right from the beg<strong>in</strong>n<strong>in</strong>g,<br />

for example P = �n + te and ||�n|| �= 1, it can be divided by P 2 = �n 2 + t(�ne +e�n) =<br />

�n 2 . A plane, as def<strong>in</strong>ed above, has to fulfill two conditions<br />

P 2 = 1 and eo · P ≤ 0.<br />

Notice that planes with d < 0 are simply oppositely oriented and occur equally<br />

frequently.<br />

Distance d and normal vector ˆn can easily be extracted<br />

d = P · e+<br />

ˆn = E · (E ∧ P).<br />

The two parameters ˆn and d appear as well <strong>in</strong> the OPNS formula (3.20): build<strong>in</strong>g<br />

the IPNS representation of the first component e ∧�a ∧�b ∧�c, only, <strong>in</strong> equation (3.20)<br />

yields<br />

(e ∧�a ∧� �<br />

b ∧�c)I = − �a · ( � �<br />

b × �c) e<br />

For the second component it is<br />

= α (de) α ∈ �.<br />

(( � b −�a) ∧ (�c −�a))E I = (�c −�a) × ( � b −�a) (3.30)<br />

= α ˆn,<br />

for the same α ∈ � as above. Its value is the magnitude of ( � b − �a) ∧ (�c − �a) and<br />

can be computed by means of equation (2.16).<br />

Hence if a plane is created by means of the OPNS representation, its orientation<br />

ˆn can be determ<strong>in</strong>ed with a hand rule 9 : the f<strong>in</strong>gers of the left hand are curled to<br />

match the direction given by the po<strong>in</strong>t sequence a-b-c; then the thumb <strong>in</strong>dicates<br />

the direction of the plane.<br />

9 Right hand sequence: a-c-b.

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

Saved successfully!

Ooh no, something went wrong!