Existence of solution to modified Gursky-Streets equation
Yi Huang, Zhenan Sui, Mingyu Xie · 2024
Read Existence of solution to modified Gursky-Streets equation by Yi Huang, Zhenan Sui, Mingyu Xie (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2412-17437.txt.
Introduction
Yi Huang, Zhenan Sui, and Mingyu Xie solve a hard geometry equation in a 2024 paper. The modified Gursky-Streets equation is fully nonlinear and arises in conformal geometry, the study of shapes under angle-preserving change. They prove solutions exist with uniform smoothness estimates in two parameter regimes. In a further case they prove existence of a weaker, viscosity-type solution that is still well behaved. The paper is a compact existence proof aimed at specialists in geometric analysis.
Catalog record
| CREATOR | Yi Huang, Zhenan Sui, Mingyu Xie |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 90 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2412-17437
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Existence of solution to modified Gursky-Streets equation Thanks: The second author is supported by National Natural Science Foundation of China (No. 12001138). Yi Huang, Zhenan Sui, Mingyu Xie Abstract . We solve the modified Gursky-Streets equation, which is a fully nonlinear equation arising in conformal geometry with uniform C 1 , 1 C^{1,1} estimates when (i) γ > 0 \gamma>0 and 1 ≤ k ≤ n 1\leq k\leq n or (ii) r > 0 r>0 and 2 s k ≤ r n 2sk\leq rn . We also prove the existence of a Lipschitz continuous viscosity solution when r ≠ 0 r\neq 0 . 2010 Mathematics Subject Classification Primary 53C21; Secondary 35J60 1. Introduction On a smooth compact Riemannian manifold (M^{n},g) of dimension n\geq 3 , we are interested in solving the following class of conformal curvature equations (1.1) u_{tt}\sigma_{k}\big(W[u]\big)-\sigma_{k}^{ij}\big(W[u]\big)u_{ti}u_{tj}=\psi(x,t)\quad\text{on}\quad M\times[0,1] subject to the boundary condition (1.2) u(\cdot,0)=u_{0},\quad u(\cdot,1)=u_{1}, where (1.3) W[u]=g^{-1}\bigg(\nabla^{2}u+sdu\otimes du+\Big(\gamma\Delta u-\frac{r}{2}|\nabla u|^{2}\Big)g+A\bigg), \gamma,s,r\in\mathbb{R} , \gamma\geq 0 , A is a smooth symmetric tensor on M with \lambda(g^{-1}A)\in\Gamma_{k} , and \psi\geq 0 is a given smooth function defined on M\times[0,1] , u_{0} , u_{1} are given smooth functions on M with \lambda\big(W[u_{0}]\big)\in\Gamma_{k} and \lambda\big(W[u_{1}]\big)\in\Gamma_{k} . Also, \nabla{u} , \nabla^{2}u and \Delta u are the gradient, Hessian and Laplace-Beltrami operator of u with respect to the background metric g respectively. In order to make the notation and computation easier, we always choose a smooth local orthonormal frame field e_{1},\ldots,e_{n} on M with respect to the metric g . Thus, u_{ti}:=\nabla_{e_{i}}u_{t} , u_{ij}:=\nabla_{e_{j}e_{i}}u , and higher order covariant derivatives can be similarly written in this manner. Besides, \sigma_{k}\big(W[u]\big):=\sigma_{k}\Big(\lambda\big(W[u]\big)\Big),\quad\sigma_{k}^{ij}\big(W[u]\big):=\frac{\partial\sigma_{k}\big(W[u]\big)}{\partial W_{ij}[u]}, where \lambda\big(W[u]\big)=(\lambda_{1},\ldots,\lambda_{n}) are the eigenvalues of the matrix W[u] , and \sigma_{k}(\lambda)=\sum\limits_{1\leq i_{1}<\cdots<i_{k}\leq n}\lambda_{i_{1}}\cdots\lambda_{i_{k}} is the k th elementary symmetric function defined on Garding’s cone \Gamma_{k}=\{\lambda\in\mathbb{R}^{n}:\sigma_{j}(\lambda)>0,\,j=1,\ldots,k\}. If we set E_{u}=u_{tt}W[u]-g^{-1}du_{t}\otimes du_{t}, we arrive at the following relation (1.4) u_{tt}\sigma_{k}\big(W[u]\big)-\sigma_{k}^{ij}\big(W[u]\big)u_{ti}u_{tj}=u_{tt}^{1-k}\sigma_{k}(E_{u}) by Proposition 2.2 . Thus equation ( 1.1 ) is equivalent to (1.5) u_{tt}^{1-k}\sigma_{k}(E_{u})=\psi(x,t). When \psi>0 throughout M\times[0,1] , following [ 16 ] we call a C^{2} function u on M\times[0,1] admissible if \lambda\big(W[u]\big)\in\Gamma_{k},\quad u_{tt}>0,\quad\sigma_{k}(E_{u})>0. Consequently, equation ( 1.1 ) or ( 1.5 ) is elliptic for C^{2} solution u with \lambda\big(W[u]\big)(x,t)\in\Gamma_{k} for any (x,t)\in M\times[0,1] (see Proposition 2.4 ). When equation ( 1.1 ) becomes degenerate, we call a C^{1,1} function u on M\times[0,1] admissible if \lambda\big(W[u]\big)\in\overline{\Gamma}_{k},\quad u_{tt}\geq 0,\quad\sigma_{k}(E_{u})\geq 0 almost everywhere. If \psi\equiv 0 , s=1 , r=1 , \gamma=0 , and A is the Schouten tensor of g , then we obtain the following degenerate conformal curvature equation (1.6) u_{tt}\sigma_{k}\big(A_{u}\big)-\sigma_{k}^{ij}\big(A_{u}\big)u_{ti}u_{tj}=0 with (1.7) A_{u}=\nabla^{2}u+du\otimes du-\frac{1}{2}|\nabla u|^{2}g+A. Equation ( 1.6 ) was introduced by Gursky and Streets [ 13 ] . In this important work, the uniqueness of the \sigma_{2} -Yamabe equation in dimension n=4 is proved. A key ingredient of their proof involves a study of equation ( 1.6 ). This equation is now known as the Gursky-Streets equation, which turns out to be the geodesic equation under a Riemannian metric defined on a conformal class of metrics. Subsequently, the Gursky-Streets equation was studied by He [ 15 ] for k=2 in dimension n\geq 4 , where he established uniform C^{1,1} regularity for k=2 and reproved the uniqueness of the \sigma_{2} equation in dimension n=4 . Then, the case when k\leq\frac{n}{2} was investigated by He, Xu and Zhang [ 16 ] by a crucial proof of the concavity of the operator associated to the equation for all 1\leq k\leq n (see Proposition 2.1 ), which relies on Garding theory of hyperbolic polynomials and results in the theory of real roots for interlacing polynomials. It is worth mentioning that the geometry of Gursky-Streets’ metric on the space of conformal metrics has a parallel theory with the geometry of the space of Kähler metrics, where the geodesic equation can be written as a homogeneous complex Monge-Ampère equation (see [ 2 , 24 , 23 , 25 , 5 , 3 , 6 ] ). Motivated by the above work, we are especially interested in the existence and regularity of solutions to the generalized Gursky-Streets equation ( 1.1 ). Under conformal deformation of metric g_{u}=e^{2u}g , the Ricci tensor \text{Ric}_{u} of g_{u} and Ric of g are related by the formula (1.8) -\frac{\text{Ric}_{u}}{n-2}=\nabla^{2}u-du\otimes du+\Big(\frac{\Delta u}{n-2}+|\nabla u|^{2}\Big)g-\frac{\text{Ric}}{n-2}. Moreover, the Schouten tensor A_{u}=\frac{1}{n-2}\Big(Ric_{u}-\frac{S_{u}}{2(n-1)}g_{u}\Big) of g_{u} and A of g are related by the formula (1.9) -A_{u}=\nabla^{2}u-du\otimes du+\frac{1}{2}|\nabla u|^{2}g-A, where S_{u} is the scalar curvature of g_{u} . If we change u into -u in ( 1.9 ), we will obtain exactly ( 1.7 ). In addition, in Gursky and Viaclovski [ 14 ] and Li and Li [ 19 ] , the following modified Schouten tensor with a parameter \tau A_{u}^{\tau}=\frac{1}{n-2}\Big(Ric_{u}-\frac{\tau S_{u}}{2(n-1)}g_{u}\Big) was introduced, which satisfies (1.10) -A_{u}^{\tau}=\nabla^{2}u-du\otimes du+\bigg(\frac{1-\tau}{n-2}\Delta u+\Big(1-\frac{\tau}{2}\Big)|\nabla u|^{2}\bigg)g-A^{\tau}. All ( 1.7 ), ( 1.9 ), ( 1.8 ) and ( 1.10 ) with \tau\leq 1 fall into the form of ( 1.3 ), which are known as conformal Schouten tensor, conformal negative Schouten tensor, conformal negative Ricci tensor and conformal negative modified Schouten tensor respectively. A central problem in conformal geometry is on the solvability of fully nonlinear Yamabe type problem which can be written in the general form (1.11) \sigma_{k}\big(W[u]\big)=e^{-2ku}\psi(x), or (1.12) \sigma_{k}\big(W[u]\big)=e^{2ku}\psi(x), where \psi(x) is a given smooth positive function defined on smooth compact manifold with or without boundary. There is a vast amount of literature on the existence and regularity of solutions to ( 1.11 ) or ( 1.12 ) ever since the work of Viaclovski [ 27 ] . In particular, we notice that the tensor ( 1.7 ), ( 1.9 ) and ( 1.8 ) may bring very different existence and regularity issues to equations in the form ( 1.11 ) or ( 1.12 ). The estimates in [ 28 ] and [ 14 ] reflect such difference (see also [ 4 ] for a more general equation). Moreover, when solving fully nonlinear Loewner-Nirenberg type problems, we observe that the negative Schouten tensor case has Lipschitz continuous solution in general, as proved by González, Li and Nguyen [ 9 ] , and there are counterexamples to C^{1} regularity given in Li and Nguyen [ 21 ] and in Li, Nguyen and Xiong [ 22 ] . In contrast, when \gamma>0 (for example, ( 1.8 )), Guan [ 10 , 11 ] obtained smooth solution to ( 1.12 ) on smooth compact manifold with boundary, based on which, smooth solution to fully nonlinear Loewner-Nirenberg problem of negative Ricci tensor was obtained in [ 26 ] . In view of these work, we are interested in whether the \Delta u term can still bring in a priori estimates up to second order for equation ( 1.1 )–( 1.2 ) when \psi>0 throughout M\times[0,1] . The following result gives an affirmative answer. Theorem 1.13 . Let (M,g) be a compact Riemannian manifold with \lambda(A)\in\Gamma_{k} and suppose that (i) \gamma>0 and 1\leq k\leq n or (ii) r>0 and 2sk\leq rn . Given a smooth positive function \psi(x,t) defined on M\times[0,1] and smooth function u_{0} , u_{1} defined on M with \lambda\big(W[u_{0}]\big)\in\Gamma_{k} and \lambda\big(W[u_{1}]\big)\in\Gamma_{k} , there exists a unique smooth solution u of ( 1.1 )–( 1.2 ) on M\times[0,1] with \lambda\big(W[u]\big)(x,t)\in\Gamma_{k} for any (x,t)\in M\times[0,1] . Moreover, we have