20.01.2015 Views

Feynman Path Integral Formulation

Feynman Path Integral Formulation

Feynman Path Integral Formulation

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.

7.4 Strong Coupling Expansion 243<br />

where, as customary, the lattice ultraviolet cutoff is set equal to one (i.e. all length<br />

scales are measured in units of the lattice cutoff). For definiteness the measure will<br />

be of the form ∫ ∫ ∞<br />

[dl 2 ]= ∏ [V d (s)] σ ∏ dlij 2 Θ[lij] 2 . (7.65)<br />

0 s<br />

ij<br />

The lattice partition function Z latt should be compared to the continuum Euclidean<br />

<strong>Feynman</strong> path integral of Eq. (2.34),<br />

∫<br />

Z cont =<br />

[dg μν ] e −λ ∫ dx √ g+ 16πG<br />

1 ∫ √ dx gR<br />

. (7.66)<br />

When doing an expansion in the kinetic term proportional to k, it will be convenient<br />

to include the λ-term in the measure. We will set therefore in this Section as in<br />

Eq. (6.93)<br />

dμ(l 2 ) ≡ [dl 2 ]e −λ 0 ∑ h V h<br />

. (7.67)<br />

It should be clear that this last expression represents a fairly non-trivial quantity,<br />

both in view of the relative complexity of the expression for the volume of a simplex,<br />

Eq. (6.5), and because of the generalized triangle inequality constraints already<br />

implicit in [dl 2 ]. But, like the continuum functional measure, it is certainly local,<br />

to the extent that each edge length appears only in the expression for the volume of<br />

those simplices which explicitly contain it. Also, we note that in general the integral<br />

∫ dμ can only be evaluated numerically; nevertheless this can be done, at least<br />

in principle, to arbitrary precision. Furthermore, λ 0 sets the overall scale and can<br />

therefore be set equal to one without any loss of generality.<br />

Thus the effective strong coupling measure of Eq. (7.67) has the properties that<br />

(a) it is local in the lattice metric of Eq. (6.3), to the same extent that the continuum<br />

measure is ultra-local, (b) it restricts all edge lengths to be positive, and (c)<br />

it imposes a soft cutoff on large simplices due to the λ 0 -term and the generalized<br />

triangle inequalities. Apart from these constraints, it does not significantly restrict<br />

the fluctuations in the lattice metric field at short distances. It will be the effect of<br />

the curvature term to restrict such fluctuation, by coupling the metric field between<br />

simplices, in the same way as the derivatives appearing in the continuum Einstein<br />

term couple the metric between infinitesimally close space-time points.<br />

As a next step, Z latt is expanded in powers of k,<br />

∫<br />

Z latt (k) = dμ(l 2 ) e k ∑ h δ h A h<br />

=<br />

∞<br />

∑<br />

n=0<br />

∫<br />

1<br />

n! kn<br />

( ) n<br />

dμ(l 2 ) ∑δ h A h . (7.68)<br />

h<br />

It is easy to show that Z(k) =∑ ∞ n=0 a n k n is analytic at k = 0, so this expansion should<br />

be well defined up to the nearest singularity in the complex k plane. A quantitative<br />

estimate for the expected location of such a singularity in the large-d limit will be<br />

given later in Sect. 7.6. Beyond this singularity Z(k) can sometimes be extended, for

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

Saved successfully!

Ooh no, something went wrong!