03.12.2012 Views

PERCOLATION AND DISORDERED SYSTEMS Geoffrey GRIMMETT

PERCOLATION AND DISORDERED SYSTEMS Geoffrey GRIMMETT

PERCOLATION AND DISORDERED SYSTEMS Geoffrey GRIMMETT

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.

222 <strong>Geoffrey</strong> Grimmett<br />

The infra-red bound emerges in the proof of (c), of which there follows an<br />

extremely brief account.<br />

We write x ⇔ y, and say that x is ‘doubly connected’ to y, if there exist<br />

two edge-disjoint open paths from x to y. We express τp(0, x) in terms of<br />

the ‘doubly connected’ probabilities δp(u, v) = Pp(u ⇔ v). In doing so, we<br />

encounter formulae involving convolutions, which may be treated by taking<br />

transforms. At the first stage, we have that<br />

�<br />

�<br />

{0 ↔ x} = {0 ⇔ x} ∪<br />

〈u,v〉<br />

� 0 ⇔ (u, v) ↔ x � �<br />

where � 0 ⇔ (u, v) ↔ x � represents the event that 〈u, v〉 is the ‘first pivotal<br />

edge’ for the event {0 ↔ x}, and that 0 is doubly connected to u. (Similar<br />

but more complicated events appear throughout the proof.) Therefore<br />

(8.27) τp(0, x) = δp(0, x) + �<br />

〈u,v〉<br />

Now, with A(0, u; v, x) = {v ↔ x off C 〈u,v〉(0)},<br />

� �<br />

Pp 0 ⇔ (u, v) ↔ x .<br />

� �<br />

Pp 0 ⇔ (u, v) ↔ x<br />

� �<br />

= pPp 0 ⇔ u, A(0, u; v, x)<br />

� � �<br />

= pδp(0, u)τp(v, x) − pEp 1 {0⇔u} τp(v, x) − 1A(0,u;v,x) �<br />

whence, by (8.27),<br />

(8.28) τp(0, x) = δp(0, x) + δp ⋆ (pI) ⋆ τp(x) − Rp,0(0, x)<br />

where ⋆ denotes convolution, I is the nearest-neighbour function I(u, v) = 1<br />

if and only if u ∼ v, and Rp,0 is a remainder.<br />

Equation (8.28) is the first step of the lace expansion, In the second step,<br />

the remainder Rp,0 is expanded similarly, and so on. Such further expansions<br />

yield the lace expansion: if p < pc then<br />

(8.29) τp(0, x) = hp,N(0, x) + hp,N ⋆ (pI) ⋆ τp(x) + (−1) N+1 Rp,N(0, x)<br />

for appropriate remainders Rp,N, and where<br />

hp,N(0, x) = δp(0, x) +<br />

N�<br />

(−1) j Πp,j(0, x)<br />

and the Πp,n are appropriate functions (see Theorem 4.2 of [180]) involving<br />

nested expectations of quantities related to ‘double connections’.<br />

j=1

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

Saved successfully!

Ooh no, something went wrong!