next up previous contents
Next: 5.6 Abstract Dimensional Spaces Up: 5. Scaling Invariance and Previous: 5.4 Algebra of Dimensional   Contents


5.5 Gauge Group Hom($ \cal {W}$, R+)

We recall that the group $ \it Hom$($ \cal {W}$,$ \bf R_{+}^{}$) is a set of homomorphisms from vector space $ \cal {W}$ (treated as an abelian group of vectors) into group $ \bf R_{+}^{}$. As there exists a natural isomorphism between the groups $ \bf R_{+}^{}$ and $ \bf R$:

exp : R $\displaystyle \ni$ y $\displaystyle \longrightarrow$ ey $\displaystyle \in$ $\displaystyle \bf R_{+}^{}$,
ln : $\displaystyle \bf R_{+}^{}$ $\displaystyle \ni$ y $\displaystyle \longrightarrow$ lny $\displaystyle \in$ $\displaystyle \bf R$,
(5.14)

it suffices to describe the structure of group $ \it Hom$($ \cal {W}$,$ \bf R$).

Let $ \left\{\vphantom{ \widehat{e}_{j} }\right.$$ \widehat{e}_{j}^{}$ $ \left.\vphantom{ \widehat{e}_{j} }\right\}$(j = 1, 2,..., n) be the vector base for $ \cal {W}$, then each element f $ \in$ $ \it Hom$($ \cal {W}$, R) is described by a single line matrix [f1,..., fn]. The value f on the vector x $ \in$ $ \cal {W}$ is described in a natural manner:

f (x) = f$\displaystyle \left(\vphantom{ \prod_{j=1}^{n} \widehat{e}_{j}^{x_{j}} }\right.$$\displaystyle \prod_{j=1}^{n}$$\displaystyle \widehat{e}_{j}^{x_{j}}$ $\displaystyle \left.\vphantom{ \prod_{j=1}^{n} \widehat{e}_{j}^{x_{j}} }\right)$ = $\displaystyle \sum_{j=1}^{n}$fjxj. (5.15)

Passing over from group R to $ \bf R_{+}^{}$ from (5.9), (5.11) we obtain

expf (x) = $\displaystyle \prod_{j=1}^{n}$efjxj. (5.16)

Defining hj = efj we obtain the element h of group $ \it Hom$($ \cal {W}$,$ \bf R_{+}^{}$) conformable to the matrix [f1,..., fn].:

h(x) = h$\displaystyle \left(\vphantom{ \prod_{j=1}^{n} \widehat{e}_{j}^{x_{j}} }\right.$$\displaystyle \prod_{j=1}^{n}$$\displaystyle \widehat{e}_{j}^{x_{j}}$ $\displaystyle \left.\vphantom{ \prod_{j=1}^{n} \widehat{e}_{j}^{x_{j}} }\right)$ = $\displaystyle \prod_{j=1}^{n}$hjxj. (5.17)

In practice the homomorphism h $ \in$ $ \it Hom$($ \cal {W}$,$ \bf R_{+}^{}$) is most often described by giving the numbers hj(j = 1, 2,..., n). The single-line matrix of homomorphism h in the base $ \widehat{e}_{j}^{}$(j = 1, 2,..., n) is thus considered to be the ordered n-ple:

[h1,..., hn], (5.18)

where hj $ \in$ $ \bf R_{+}^{}$ for j = 1, 2,..., n is obvious since (5.14), (5.15) are continuous mappings, the homomorphisms (5.17) are also continuous.

The here described procedure of choosing base $ \left\{\vphantom{ \widehat{e}_{j} }\right.$$ \widehat{e}_{j}^{}$ $ \left.\vphantom{ \widehat{e}_{j} }\right\}$ and the parameterization of the group $ \it Hom$($ \cal {W}$,$ \bf R_{+}^{}$) by matrices (5.18) are denoted as the choice of a dimensional base. This method has been in general use.

The gauge group does not depend either on the dimensional space construction. To describe $ \it Hom$($ \cal {W}$,$ \bf R_{+}^{}$) it suffices to give the space of pure $ \cal {W}$ dimensions. The gauge group is, therefore, a rather universal object.

Let r $ \in$ $ \it Hom$($ \cal {W}$,$ \bf R_{+}^{}$) be the fixed homomorphism of groups. The kernel of homomorphism r is defined and denoted:

Ker  r = {$\displaystyle \omega$ $\displaystyle \in$ $\displaystyle \cal {W}$ : r($\displaystyle \omega$) = 1}.

Of course, Ker  r is a subgroup of $ \cal {W}$ as a kernel of the homomorphism of groups, moreover it is also a vector subspace.

Let the two different homomorphisms r1, r2 $ \in$ $ \it Hom$($ \cal {W}$,$ \bf R_{+}^{}$) have the same kernel Ker  r1 = Ker  r2. It is then easy to verify that one of them is the power of the other, i.e., there exists the number $ \alpha$ $ \in$ R for which r2($ \omega$) = r1$\scriptstyle \alpha$($ \omega$) (for $ \omega$ $ \in$ $ \cal {W}$). Let $ \omega$ $ \in$ $ \cal {W}$ = $ \cal {W}$0 $ \oplus$ Ker  r1, then there exists a unique decomposition of the vector

$\displaystyle \omega$ = $\displaystyle \omega_{0}^{}$v,

where $ \omega_{0}^{}$ $ \in$ $ \cal {W}$0, v $ \in$ Ker  r1. As dim$ \cal {W}$0 = 1 each vector $ \omega$ has the representation

$\displaystyle \omega$ = $\displaystyle \sigma^{a}_{0}$v,

where $ \sigma_{0}^{}$ is the fixed vector from $ \cal {W}$0, a $ \in$ R and v $ \in$ Ker  r1. It is now easy to compare the values of the two homomorphisms:

r1($\displaystyle \omega$) = r1a($\displaystyle \sigma_{0}^{}$),      r2($\displaystyle \omega$) = r2a($\displaystyle \sigma_{0}^{}$). (5.19)

Since there always exists such a number $ \alpha$ $ \in$ R, that r2($ \sigma_{0}^{}$) = r1$\scriptstyle \alpha$($ \sigma_{0}^{}$), we obtain from (5.19)

$\displaystyle \forall$$\displaystyle \omega$ $\displaystyle \in$ $\displaystyle \cal {W}$        r2($\displaystyle \omega$) = r1$\scriptstyle \alpha$($\displaystyle \omega$).

Let us note at the same time that the equalities (5.19) are equivalent to condition

$\displaystyle \forall$r $\displaystyle \in$ $\displaystyle \it Hom$($\displaystyle \cal {W}$,$\displaystyle \bf R_{+}^{}$)    $\displaystyle \forall$$\displaystyle \omega$ $\displaystyle \in$ $\displaystyle \cal {W}$    $\displaystyle \forall$a $\displaystyle \in$ R      r($\displaystyle \omega^{a}_{}$) = ra($\displaystyle \omega$), (5.20)

Thus (5.19) shows that the elements of the gauge group are in fact linear transformations and not merely group homomorphisms. Obtained results are presented as a theorem.

Theorem 5.1.   The gauge group $ \it Hom$($ \cal {W}$,$ \bf R_{+}^{}$) is a vector space over R (in a multiplicative notation). The unit of space $ \it Hom$($ \cal {W}$,$ \bf R_{+}^{}$) is the homomorphism e:

$\displaystyle \forall$$\displaystyle \omega$ $\displaystyle \in$ $\displaystyle \cal {W}$      e($\displaystyle \omega$) = 1,

the multiplication by a scalar are the powers:

$\displaystyle \forall$$\displaystyle \alpha$ $\displaystyle \in$ R    $\displaystyle \forall$r $\displaystyle \in$ $\displaystyle \it Hom$($\displaystyle \cal {W}$,$\displaystyle \bf R_{+}^{}$)        (r$\scriptstyle \alpha$)($\displaystyle \omega$) = r$\scriptstyle \alpha$($\displaystyle \omega$),

the vector product - the group product

$\displaystyle \forall$r1, r2 $\displaystyle \in$ $\displaystyle \it Hom$($\displaystyle \cal {W}$,$\displaystyle \bf R_{+}^{}$)    $\displaystyle \forall$$\displaystyle \omega$ $\displaystyle \in$ $\displaystyle \cal {W}$  (r1r2)($\displaystyle \omega$) = r1($\displaystyle \omega$)r2($\displaystyle \omega$).

The dimension of the vector space $ \it Hom$($ \cal {W}$,$ \bf R_{+}^{}$) is identical with the dimension of $ \cal {W}$ :

dim$\displaystyle \it Hom$($\displaystyle \cal {W}$,$\displaystyle \bf R_{+}^{}$) = dim$\displaystyle \cal {W}$ (5.21)

Proof:
The fact that $ \it Hom$($ \cal {W}$,$ \bf R_{+}^{}$) is a vector space results from (5.20) and the Definition 5.9. The equality of dimensions (5.21) results from the parameterization by matrices (5.18).

Finally, let us note that the gauge group is a rather non-typical object from the point of view of classical dimensional analysis. Let r $ \in$ $ \it Hom$($ \cal {W}$,$ \bf R_{+}^{}$), then r : $ \cal {W}$ $ \longrightarrow$ $ \bf R_{+}^{}$ so that the arguments of homomorphism r are quatities of a dimensional type, e.g., mass, force, charge, while the values are numbers from $ \bf R_{+}^{}$. Thus we have mappings of dimensions into dimensionless numbers.


next up previous contents
Next: 5.6 Abstract Dimensional Spaces Up: 5. Scaling Invariance and Previous: 5.4 Algebra of Dimensional   Contents
Wojciech Myszka
1999-09-22