Lipschitz regularity of harmonic map heat flows into C A T ( 0 ) CAT(0) spaces
Hui-Chun Zhang, Xi-Ping Zhu · 2026
Read Lipschitz regularity of harmonic map heat flows into C A T ( 0 ) CAT(0) spaces by Hui-Chun Zhang, Xi-Ping Zhu (2026) 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-2601-20579.txt.
Introduction
Hui-Chun Zhang and Xi-Ping Zhu study a heat flow for harmonic maps in this 2026 math paper. A harmonic map is a map between shapes that stretches them as little as possible. The flow slowly changes such a map over time, much like heat spreads out. Here the maps land in CAT(0) spaces, a broad class of spaces with no positive curve.
The authors trace the line of work behind their question. In 1964, Eells and Sampson proved long-time existence and convergence for flows into spaces that do not curve up. Hamilton later took up the same flow with boundary conditions added. In 1992, Gromov and Schoen built a theory of such maps into CAT(0) spaces. Zhang and Zhu ask how smooth, in the Lipschitz sense, these flows stay.
Catalog record
| CREATOR | Hui-Chun Zhang, Xi-Ping Zhu |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 138 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-20579
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Lipschitz regularity of harmonic map heat flows into C A T ( 0 ) CAT(0) spaces Hui-Chun Zhang, Xi-Ping Zhu Abstract . In 1964, Eells and Sampson proved the famous long-time existence and convergence for the harmonic map heat flow into non‑positively curved Riemannian manifolds. Subsequently, Hamilton investigated the corresponding initial-boundary problem. In 1992, Gromov and Schoen developed a variational theory of harmonic maps into C A T ( 0 ) CAT(0) metric spaces. This progress naturally motivated the study of the harmonic map heat flow into singular metric spaces. In the 1990s, Mayer and Jost independently studied convex functionals on C A T ( 0 ) CAT(0) spaces and extended Crandall-Liggett’s theory of gradient flows from Banach spaces to C A T ( 0 ) CAT(0) spaces to obtain the weak solutions, called semi-group weak solutions, for the harmonic map heat flow into C A T ( 0 ) CAT(0) spaces. Very recently, Lin, Segatti, Sire, and Wang used an elliptic approximation method to obtain another class of weak solutions, called suitable weak solutions in the sense of the Evolution Variational Inequality (EVI), to the harmonic map heat flow into C A T ( 0 ) CAT(0) spaces. They proved that these solutions are Lipschitz in space and 1 2 \frac{1}{2} -Hölder continuous in time. Since the semi-group weak solutions of the harmonic map heat flow enjoy the favorable long-time existence, uniqueness and well established long-time behaviors, it is natural to ask if the semi-group solutions possess the Lipschitz regularity. In the present paper, we answer this question. We show that the semi‑group weak solutions of the harmonic map heat flow into CAT(0) spaces are Lipschitz continuous in both space and time. We also establish an Eells–Sampson‑type Bochner inequality. 1. Introduction The harmonic map heat flow deforms a given map u_{0}:M\to N\subset\mathbb{R}^{\ell} between two compact Riemannian manifolds via the equation (1.1) \begin{cases}\partial_{t}u&=\Delta u+A(u)(\nabla u,\nabla u),\quad(x,t)\in M\times(0,+\infty),\\ u|_{t=0}&=u_{0},\quad x\in M,\end{cases} where \Delta is the Laplace-Beltrami operator on M and A is the second fundamental form of the embedding N\subset\mathbb{R}^{\ell}. In 1964, Al’ber [ 1 , 2 ] introduced the flow, Eells and Sampson [ 16 ] proved the following famous theorem. Theorem 1.1 (Eells-Sampson [ 16 ] ) . Let M,N be two compact Riemannian manifolds without boundaries. Suppose the sectional curvature of N is non-positive. Then for any u_{0}\in C^{\infty}(M,N) , the flow ( 1.1 ) admits a unique, smooth solution u\in C^{\infty}(M\times[0,+\infty),N) . Moreover, there exists a subsequence of \{u(\cdot,t_{j})\} which converges to a harmonic map, as t_{j}\to+\infty . This smoothness of u(x,t) in Theorem 1.1 is important, because it makes t\mapsto u(\cdot,t) a continuous deformation that preserves the homotopy class of the initial value u_{0} . When M has non-empty smooth boundary \partial M , Hamilton [ 25 ] studied the initial-boudnary problem (1.2) \begin{cases}\partial_{t}u&=\Delta u+A(u)(\nabla u,\nabla u),\quad(x,t)\in M\times(0,+\infty),\\ u|_{t=0}&=u_{0},\quad x\in M,\\ u(x,t)&=\phi(x),\quad x\in\partial M,\ \ t>0.\end{cases} Hamilton proved that if N has non-positive sectional curvature, then for given u_{0},\phi\in C^{\infty}(\overline{M},N) , there exists a unique smooth solution in C^{\infty}(\overline{M}\times[0,+\infty),N). In 1992, Gromov and Schoen [ 22 ] developed a theory of harmonic maps into CAT(0) complexes, by using the calculus of variations, to establish the p -adic superrigidity of lattices in groups of rank one. After this, many interesting results for harmonic maps into or between singular metric spaces have been obtained (see, e.g., [ 36 , 8 , 37 , 31 , 32 , 33 , 34 , 40 , 15 , 53 , 51 , 52 , 44 , 10 , 12 , 29 , 56 , 55 , 17 , 6 , 18 , 23 , 45 , 19 , 57 ] and references therein). Remark that variational methods have their limitations; for instance, the convergence in W^{1,2}(M,N) does not preserve some topological properties (examples can be found in [ 42 , page 110] ). This naturally motivates the study of the harmonic map heat flow into singular metric spaces. 1.1. Harmonic map heat flows into CAT(0) spaces Since a general CAT(0) space Y might not be locally compact and can not be embedded into an Euclidean space, one has to use an intrinsic approach to study the harmonic maps and their heat flows into CAT(0) spaces. The theory of Sobolev spaces for maps into metric spaces has been well-developed [ 36 , 31 , 35 , 39 , 28 , 47 , 21 ] . Let u be a map from a bounded open domain \Omega\subset M to an arbitrary metric space (Y,d_{Y}) . It is called a L^{2}(\Omega,Y) map if its range is separable and if for some P\in Y , the function d_{Y}(P,u(\cdot)) is in L^{2}(\Omega) . Given a map u\in L^{2}(\Omega,Y), for each \epsilon>0 , the approximating energy E^{u}_{\epsilon} is defined as a functional on C_{c}(\Omega) , the space of continuous functions compactly supported on \Omega : E^{u}_{\epsilon}(\phi):=\int_{\Omega}\phi(x)e^{u}_{\epsilon}(x){\rm d}\mu(x),\quad e^{u}_{\epsilon}(x):=\frac{n(n+2)}{\omega_{n-1}}\int_{B_{\epsilon}(x)\cap\Omega}\!\!\frac{d^{2}_{X}\big(u(x),u(y)\big)}{\epsilon^{n+2}}{\rm d}\mu(y) for all \phi\in C_{c}(\Omega) , where \omega_{n-1} is the volume of (n-1) -dimensional sphere \mathbb{S}^{n-1} with the standard metric. In [ 36 ] , Korevaar-Schoen proved that (1.3) \lim_{\epsilon\to 0^{+}}E^{u}_{\epsilon}(\phi)=E[u](\phi) for some positive functional E[u](\phi) on C_{c}(\Omega) . A L^{2}(\Omega,Y) -map u is called in W^{1,2}(\Omega,Y) if \sup_{\phi\in C_{c}(\Omega),\ 0\leqslant\phi\leqslant 1}E[u](\phi)<+\infty. Let (Y,d_{Y}) be a CAT(0) space (i.e., a globally non-positively curved metric space in the sense of Alexandrov). The functional u\mapsto E[u] is convex on the metric space L^{2}(\Omega,Y) , which is allowed to take the value +\infty . In the 1990’s, Jost [ 32 ] and Mayer [ 43 ] independently studied its resolvent J_{h} , that is, for any given h>0 and u_{0},\psi\in W^{1,2}(\Omega,Y) , J_{h}(u_{0}) is the (unique) minimizer of u\mapsto\frac{1}{2}E[u]+\frac{1}{2h}D^{2}(u,u_{0}),\qquad D^{2}(u,u_{0}):=\int_{\Omega}d^{2}_{Y}(u(x),u_{0}(x)){\rm d}\mu(x), in the space W^{1,2}_{\psi}(\Omega,Y):=\{v\in W^{1,2}(\Omega,Y)|\ d_{Y}(v(x),\psi(x))\in W^{1,2}_{0}(\Omega)\} . The first theory of harmonic map heat flow into CAT(0) space was established by Mayer [ 43 ] . He extended Crandall-Liggett’s theory of gradient flows from Banach spaces to CAT(0) spaces, proving that the limit (1.4) u(x,t)=F_{t}(u_{0}):=\lim_{m\to+\infty}J^{m}_{t/m}(u_{0}) exists for every t\in(0,+\infty) , and satisfies the semi-group property F_{t+s}(u_{0})=F_{t}\circ F_{s}(u_{0}) for all t,s>0 . The resulting function t\mapsto F_{t}(u_{0})\in L^{2}(\Omega,Y) is called a semi-group weak solution to the harmonic map heat flow on \Omega with initial u_{0} and boundary data \psi . About the same time, Jost [ 35 ] also obtained the long-time existence of semi-group weak solutions. Furthermore, Mayer [ 43 ] established the uniqueness of the semi-group weak solutions in W^{1,2}(\Omega,Y) , and the regularity (1.5) t\mapsto u(x,t)\in Lip\big((t_{0},+\infty),L^{2}(\Omega,X)\big)\cap C^{1/2}\big([0,T),L^{2}(\Omega,X)\big) for any t_{0}>0 and T<+\infty . Moreover, the long-time asymptotic limit u_{\infty}:=\lim_{t\to+\infty}F_{t}(u_{0}) exists, and is the unique harmonic map with boundary data \psi (i.e., the minimizer of E[u] in W^{1,2}_{\psi}(\Omega,Y) ). In [ 3 ] , Ambrosio, Gigli and Savaré introduced a new idea to study the gradient flows of convex functions on a non-smooth metric space via the Evolution Variational Inequality (EVI), \frac{1}{2}\frac{\rm d}{\rm dt}D^{2}(u(t),v)+E[u(t)]\leqslant E[v]\quad{\rm in}\ \ \mathscr{D}^{\prime}(0,+\infty),\quad\forall\ v\in W^{1,2}. Moreover, Gigli and Nobili [ 20 ] showed recently the existence and uniqueness of the gradient flows in the sense of the EVI, and some behaviors similar to what happens for gradient flows in Hilbert spaces. Very recently, Lin, Segatti,