# Variational principle (in gravity)
## Related
- [[0102 Hayward term]]
- [[0006 Higher-derivative gravity]]
- [[0047 Renyi at finite n for higher derivative gravity]]
- [[0127 Black hole thermodynamics]]
## Refs
- [[2008#Dyer, Hinterbichler]]
- [[SmolicTaylor2013]]
- [[MadsonBarrow1989]]
- [[GuarnizoCastanedaTejeiro2010]] f(R)
- [[BuenoCanoLassoRamirez2016]] f(Lovelock)
- [[TeimouriTalaganisEdholmMazumdar2016]]
- [[BuenoCanoRuiperez2018]] Einstein cubic gravity
- [[2022#Erdmenger, Hess, Matthaiakakis, Meyer]]: general prescription
- in AdS
- [[2005#Hollands, Ishibashi, Marolf]]
- [[HohmTonni2010]]: arXiv:1001.3598
- [[2023#Cassani, Ruiperez, Turetta]] gauged [[0332 Supergravity|SUGRA]]
## Why boundary term
- in Hamiltonian method, it is needed to give the right ADM energy [[1995#Hawking, Horowitz]]
- ensure correct composition properties of PI
- see [[GibbonsHawking1977]]
- semiclassical BH entropy in Euclidean calculation comes solely from this term
- see [[1992#Brown, York (a)]]
## Einstein
- $S=S_{\mathrm{EH}}+S_{\mathrm{GHY}} =\frac{1}{16\pi G} \int d^{4} x \sqrt{-g} R+\frac{1} {8\pi G}\oint d^{3} x \sqrt{|h|} K$
- Neumman BC: [[KrishnanRaju2016]][](https://arxiv.org/pdf/1605.01603.pdf)
- Robin BC: [[KrishnanMaheshwariSubramanian2017]][](https://arxiv.org/pdf/1702.01429.pdf)
- separation of bulk and boundary action: [[Bhattacharya2022]][](https://arxiv.org/pdf/2207.08199.pdf)
## $f(R)$
- [[2008#Dyer, Hinterbichler]]
- $S\sim\int d^{4} x \sqrt{|g|} F(R)+2 \oint d^{3} x \sqrt{|h|} F^{\prime}(R) K$
- straightforward generalisation of Einstein
- boundary term needed to make the correspondence to scalar-tensor theory hold even at the boundary
- see [[0006 Higher-derivative gravity]]
- [[2019#Harlow, Wu]]
## $f(Ricci)$
- [[WangZhao2018]]@[](https://arxiv.org/pdf/1812.01854.pdf)
- $\mathcal{L}_{B}=f^{\prime \mu \nu}\left(K_{\mu \nu}+n_{\mu} n_{\nu} K\right)$
- using $I_{b u l k}=\int_{\mathcal{M}} \mathrm{d}^{n} x \sqrt{g}\left[f\left(\phi_{\mu \nu \rho \sigma}, g_{\mu \nu}\right)-\psi^{\mu \nu \rho \sigma}\left(\phi_{\mu \nu \rho \sigma}-R_{\mu \nu \rho \sigma}\right)\right]$
- or $\mathcal{L}_{B}=n_{\mu}\left(f^{\prime \rho \mu} \Gamma_{\nu \rho}^{\nu}-f^{\prime \rho \nu} \Gamma_{\rho \nu}^{\mu}\right)$
- using $I_{b u l k}=\int_{\mathcal{M}} \mathrm{d}^{n} x \sqrt{g}\left[f\left(\phi_{\mu \nu}, g_{\mu \nu}\right)-\psi^{\mu \nu}\left(\phi_{\mu \nu}-R_{\mu \nu}\right)\right]$
## $f(Riem)$
- boundary action $\bar{S}=-\oint_{\partial \mathcal{M}} \mathrm{d} \Sigma_{\mu} n^{\mu} \Psi \cdot K$
- n.b. [[2009#Deruelle, Sasaki, Sendouda, Yamauchi]] defines $d\Sigma_\mu$ w.r.t. inward normal for timelike and outward for spacelike, whereas [[2018#Jiang, Zhang]] uses outward normal for all cases
## GB
- for 4-deriv theories, the only combination that allows good variational principle with one BC is GB
- shown in [[Bunch1981]]
- see also [[0341 Lovelock gravity]] and [[0425 Gauss-Bonnet gravity]]
## Theorems and general understanding
- [[Gieres2022]][](https://arxiv.org/abs/2205.01459): Improvement of a conserved current density v.s. adding a total derivative to a Lagrangian density