Structure of Exact Renormalization Group Equations for field theory

TL;DR

This paper demonstrates a Legendre transformation linking the exact renormalization group equations for the full action and effective action, removing the need for explicit cutoff procedures and clarifying their theoretical structure.

Contribution

It establishes a simple Legendre transformation relation between the scale-dependent full action and effective action in the RG framework, avoiding explicit cutoff references.

Findings

The Legendre transformation connects $S[\, ext{phi} ext{,}t]$ and $\, ext{ extGamma}[ ext{ extPhi} ext{,}t]$ without explicit cutoffs.

The approach aligns with dimensional regularization, using scale rather than cutoff for renormalization.

The method preserves properties of fixed points and simplifies the RG equations in field theory.

Abstract

It is shown that exact renormalization group (RG) equations (including rescaling and field-renormalization) for respectively the scale-dependent full action $S[\phi,t]$ and the scale-dependent full effective action $\Gamma[\Phi,t]$ --in which $t$ is the "RG-time" defined as the logarithm of a running momentum scale-- may be linked together by a Legendre transformation as simple as $\Gamma[\Phi,t] -S[\phi,t] + \phi \cdot \Phi=0$, with $\Phi(x) =\delta S[\phi] /\delta \phi(x) $ (resp. $\phi(x) =-\delta \Gamma [\Phi]/\delta \Phi(x) $), where $\phi$ and $\Phi$ are dimensionless-renormalized quantities. This result, in which any explicit reference to a "cutoff procedure" is absent, makes sense in the framework of field theory. It may be compared to the dimensional regularization of the perturbative field theory, in which the running momentum scale is a pure scale of reference and not a momentum cutoff. It is built from the Wilson historic first exact RG equation in which the field-renormalization step is realized via an operator which is redundant and exactly marginal at a fixed point, the properties of which are conserved by the Legendre transformation and which modifies the usual removal of the overall UV cutoff $\Lambda_{0}$ by associating it with the removal of an overall IR cutoff $\mu $. Because the final equations do not refer to any true cutoff (even for the scale-dependent $\Gamma $), it reinforces the idea that one may get rid of the achronistic procedure of "regularizing" the theory via an explicit "cutoff function", procedure which is often seen as an inconvenience to treat modern problems in field theory.