---
title: RG-Improved Einstein-Hilbert Action
url: https://www.emergentmind.com/topics/renormalization-group-improved-einstein-hilbert-action
type: topic
---

# RG-Improved Einstein-Hilbert Action

The Renormalization-Group Improved Einstein-Hilbert Action denotes a class of effective gravitational actions in which the Newton constant $G$ and the cosmological constant $\Lambda$ are replaced by dynamically running, scale-dependent quantities $G(k)$ and $\Lambda(k)$—or, in general, by functionals $G(x)$ and $\Lambda(x)$ encoding quantum corrections arising from coarse-graining at variable energy scales. This construction, motivated by the Asymptotic Safety scenario, functional renormalization group (FRG) equations, and various coarse-graining frameworks, captures the leading quantum-gravitational effects at the level of the action, and provides a technically controlled method of extrapolating semiclassical general relativity beyond its naive domain of validity [1710.01972][2411.02480][2412.14108][2506.03601].

## 1. Functional Renormalization Group and the Einstein-Hilbert Truncation

The foundational step for RG improvements in gravity is the functional renormalization group equation for the effective average action $\Gamma_k[g]$, typically implemented using the Wetterich equation or background field flow:

$$
\partial_t \Gamma_k[g] = \frac12 \text{Tr}\left[ \left(\Gamma_k^{(2)} + \mathcal{R}_k \right)^{-1}\, \partial_t \mathcal{R}_k \right]_{\text{hh}}
- \text{Tr} \left[ \left(\Gamma_{k,gh}^{(2)} + \mathcal{R}_k^{gh} \right)^{-1}\, \partial_t \mathcal{R}_k^{gh} \right],
$$

where $k$ is the IR cutoff scale, $\mathcal{R}_k$ is a regulator, and $\Gamma_k^{(2)}$ denotes the Hessian w.r.t. metric fluctuations [2411.02480][1710.01972]. The standard approximation (Einstein-Hilbert truncation) takes

$$
\Gamma_k[g] = \frac{1}{16\pi G_k} \int d^4x \sqrt{g} \left(-R + 2\Lambda_k \right)
$$

with $G_k$ and $\Lambda_k$ running according to RG flow equations. The flows are often recast in terms of dimensionless variables $g(k) = k^2 G_k$ and $\lambda(k) = \Lambda_k / k^2$.

For instance, using an optimized Litim regulator [1710.01972][2411.02480]:

\[
\begin{aligned}
\partial_t g &= (2 + \eta_N) g \\
\partial_t \lambda &= (\eta_N - 2)\lambda + \text{thresholds}(g, \lambda) \\
\eta_N &\equiv -\partial_t \ln G_k,
\end{aligned}
\]

with $\eta_N$ (anomalous dimension) and threshold functions determined by the heat kernel expansion and field content. In pure gravity, a non-Gaussian fixed point (NGFP) is often found, $g_* = \lim_{k\to\infty} g(k)$, $\lambda_* = \lim_{k\to\infty} \lambda(k)$, which is crucial for ultraviolet completeness in the Asymptotic Safety approach [1710.01972][1012.4280][2506.03601].

## 2. RG Improvement Prescription and Scale Setting

The construction of the RG-improved action requires relating the unphysical parameter $k$ to spacetime properties. This is implemented by promoting $k \rightarrow k(x)$, turning scale-dependent couplings into spacetime-dependent functions:

$$
S_\text{imp}[g] = \frac{1}{16\pi G(k(x))} \int d^4x\, \sqrt{g} \left( -R + 2\Lambda(k(x)) \right).
$$

The $k(x)$ identification is scenario-dependent:

- In FLRW cosmology: $k \sim \xi / a(t)$ or $k \sim \xi / t$
- In curvature-based settings: $k^2 \sim \xi R(x)$
- For black holes: $k^2(r; M) \sim G_0 M / r^3$ [2601.17115]

The self-consistent or variational scale-setting can also be enforced by extremizing the action w.r.t. $k(x)$, resulting in a condition, e.g.,

$$
R \frac{d}{dk} \left( \frac{1}{G(k)} \right) - 2 \frac{d\Lambda(k)}{dk} = 0,
$$

which gives a functional expression $k^2 = k^2(R, ...)$ and yields an $f(R)$, $f(R,\phi)$, or scalar-tensor action [1204.1483][1504.05119].

## 3. Structure and Generalization of the RG-Improved Action

The RG-improved Einstein-Hilbert action generically takes the form

$$
\Gamma_\text{imp}[g] = \frac{1}{16\pi G(x)} \int d^4x \sqrt{-g} [-R + 2\Lambda(x)],
$$

where $G(x)$ and $\Lambda(x)$ are functionals of the metric through their dependence on $k(x)$. This construction can be interpreted as a specific scalar-tensor theory with vanishing scalar kinetic term, as seen by defining $\phi(x) = 1/G(x)$ [1504.05119]. In this scalar-tensor embedding, the Lagrangian is

$$
(16\pi)^{-1} [ \phi R - 2\phi V\{\phi\} ],
$$

where $V\{\phi\}$ is derived from $\Lambda(k)$ once $k=k(\phi)$ is inverted.

In the context of spherically symmetric reduction or in effective two-dimensional (Horndeski) dilaton gravity [2601.17115], RG-improvement modifies the dilaton coupling functions but preserves the second-order character of the field equations.

Partition function normalization has been addressed to remove vacuum energy divergences, ensuring the RG-improved action approaches the standard Einstein-Hilbert term in the IR [2411.02480]. Real-space RG approaches with curvature coarse-graining show that higher-order corrections (quadratic and beyond) are systematically generated at successive orders of block averaging [1610.06143].

## 4. Field Equations, Consistency Relations, and Canonical Structure

Variation of the RG-improved action with respect to $g_{\mu\nu}$ yields modified Einstein equations with extra terms due to gradients of the running couplings:

\[
\frac{\delta S_\text{imp}}{\delta g^{\mu\nu}}
= \frac{1}{16\pi G(x)} (R_{\mu\nu} - \frac{1}{2} R g_{\mu\nu})
+ \frac{\Lambda(x)}{8\pi G(x)} g_{\mu\nu}
- \frac{1}{16\pi} \left[ \nabla_\mu \nabla_\nu - g_{\mu\nu} \Box \right] (G^{-1}(x)) + \cdots
\]

Energy-momentum conservation is ensured if the consistency (integrability) condition

$$
\partial_\alpha(\Lambda(x)/G(x)) = \frac{1}{2} R\, \partial_\alpha G^{-1}(x)
$$

holds [1504.05119][1204.1483]. In the canonical ADM or Hamiltonian framework, treating $G(x)$ and $\Lambda(x)$ as external fields modifies the constraint algebra, generally rendering it second class off minisuperspace and breaking full spacetime diffeomorphism invariance unless special gauge choices are made [1902.03014]. The comparison with related Brans–Dicke-type theories reveals an inequivalent constraint structure, despite superficial similarity at the level of field content.

## 5. Cosmological and Black-Hole Applications

Implementation of the RG-improved action in FLRW cosmology modifies the Friedmann equations via directly time-dependent $G(t)$ and $\Lambda(t)$, yielding

$$
H^2 = \frac{8\pi}{3} G(t) \rho + \frac{\Lambda(t)}{3} + \text{RG-induced terms},
$$

where RG-induced terms can lead to singularity resolution, cosmological bounces, or inflationary-like expansion without explicit inflaton fields [1710.01972][1204.1483]. The specific UV scaling $G(t) \sim a(t)^2$, $\Lambda(t) \sim 1/a(t)^2$ near the NGFP softens the big bang singularity [1710.01972]. Causal Dynamical Triangulations volume profiles are closely reproduced by the NGFP-dominated RG-improved cosmology.

For black holes, the improved Schwarzschild metric with $G(r)$ regularizes the central singularity given suitable scale identification $k(r)$, as demonstrated by explicit construction in 2D Horndeski reduction [2601.17115]. Multiple types of improvement exist (action, field equation, or solution-level), agreeing at large $r$ but diverging in the near-horizon/core regime.

## 6. Extensions: Higher-Derivative Gravity, Extra Dimensions, and Matter Couplings

The RG-improved approach generalizes to truncations including $R^2$, $R_{\mu\nu}^2$, or scalar-matter couplings [2506.03601][1204.1483][1012.4280]. At the NGFP, the action flows to a scale-invariant $R^2$ theory—a universal attractor in many truncations [1204.1483]. In models with matter or higher-dimensional backgrounds, the flow equations for $G_k$ and $\Lambda_k$ acquire threshold corrections and show dimensional crossovers (e.g., $D=4\to5$, with $G_k \sim g_*\, k^{2-D}$ and corresponding dimensional reduction when extra dimensions compactify) [1809.11018].

The essential RG approach removes inessential (field-redefinable) couplings to distill the flow to the physically relevant ones—here, just Newton’s constant and vacuum energy—yielding simple RG-improved Einstein–Hilbert actions with running couplings [2506.03601]. In scalar–tensor and nonminimal coupling scenarios, proper account of frame transformations and gravitational contact terms is essential to avoid apparent anomalies and ensure frame invariance of the RG flow [2210.15640].

## 7. Physical Consequences, Universality, and Open Issues

The RG-improved Einstein–Hilbert action implements quantum-corrected dynamics with potentially far-reaching implications:

- **Singularity resolution:** The NGFP scaling softens or eliminates cosmological and black-hole singularities [1710.01972][1204.1483][2601.17115].
- **Cosmic acceleration:** Running couplings can mimic dark energy, induce inflation, or modify structure formation [1504.05119][1204.1483].
- **Dark matter phenomenology:** System-dependent RG scales permit fitting of galactic rotation curves and cluster profiles without additional matter [1504.05119].
- **Modified gravity models:** RG improvement provides a systematic path to $f(R)$-type, scalar–tensor, or Horndeski theories with controlled quantum origin [1204.1483][2601.17115].
- **Constraint structure and covariance:** The explicit dependence of $G(x), \Lambda(x)$ may break diffeomorphism invariance in the full canonical theory; suitable gauge or model reductions restore tractability [1902.03014].

Central open issues include the robustness of the fixed point structure to truncation scheme and measure definitions [2412.14108], the detailed mapping from field theory RG-scale to geometric invariants [1710.01972][2506.03601], and the quantitative physical predictions (e.g., in primordial power spectra or black hole thermodynamics) subject to experimental verification.

---

**Key Papers**: [1710.01972], [2411.02480], [2412.14108], [2506.03601], [1204.1483], [1504.05119], [2601.17115], [1902.03014], [1012.4280], [1610.06143], [1809.11018], [2210.15640].

Source: https://www.emergentmind.com/topics/renormalization-group-improved-einstein-hilbert-action