Green's function for a hierarchical self-avoiding walk in four dimensions (Q1416911): Difference between revisions

The paper is the second in a series of two papers in which the asymptotic end-to-end distance of a weakly self-avoiding walk on the four-dimensional hierarchical lattice \(G\) is studied. The first paper is [\textit{D. C. Brydges} and \textit{J. Z. Imbrie}, End-to-end distance from the Green's function for a hierarchical self-avoiding walk in four dimensions. Commun. Math. Phys. 239, No. 3, 523--547 (2003; Zbl 1036.82011)]. The interacting Green's function is defined by the Laplace transform \[ G_{\lambda}(\beta,x)=\int^{\infty}_0e^{-\beta T}{\mathbf E_0\bigg({\mathbf 1}_{\omega(T)=x}e^{-\lambda\int_G\tau^2_xdx} \bigg)}\,dT \] where \(\tau_x=\tau_x^{(T)}\) is the time up to \(T\) that \(\omega(t)\) is at site \(x\). The main result of the paper says that the interacting Green's function for the Markov process, \(\lambda=0\), but at a different value of the \(\beta\) parameter which depends on \(x\) and \((\beta,\lambda)\) and which can be accurately calculated when \(\lambda\) is small. The error in the approximation decays more rapidly than the Green's function because it contains, as a prefactor, the ''running coupling constant''. Also the Green's function is analytic in \(\beta\) in a sector in the complex plane with opening angle greater than \(\pi .\) The paper also completes the proof that the expected value grows as a constant times \(\sqrt{T}\log^{1/8}\big(1+O\big({\log\log T \over \log T}\big)\big).\)