Liouville theorems for harmonic map heat flow along ancient super Ricci flow via reduced geometry (Q2050163): Difference between revisions

This is a quite interesting article, in which Liouville type theorems are proved for the harmonic map heat flow from a time-dependent Riemannian manifold, ancient super-solution of the Ricci flow, i.e. satisfying \[ \partial_t g \geq -2 \mathrm{Ricci} \quad (t\in ]-\infty , 0]). \] As to the target, it is just assumed to be complete and simply-connected with an upper bound on its sectional curvature. There are two distinct sets of conditions required to conclude that a solution of the backward harmonic map flow must be constant (``backward'' because one of the first things the authors do is to reverse the time parameter), expressed in terms of Perelman's reduced geometry. i) On the metric flow: admissible (so that the reduced distance is achieved), completeness and more technical conditions (though the significance of the bounds they imply are very clear) on the Müller and trace Harnack quantities. ii) On the map itself, assumptions are made (in both main theorems) on the asymptotic behaviour of the distance function of the image compared with the reduced distance on the domain. If, in the first theorem, the target manifold's sectional curvature is assumed to be negative, this condition is then relaxed to a uniform upper bound at the cost of an extra assumption on the image being in a sufficiently small geodesic ball. The article concludes with a nice discussion of the sharpness of an improved static version of the previous theorem and the Schoen-Uhlenbeck smooth radial harmonic maps from \(\mathbb R^m\) to \(\mathbb S_{+}^m\) for \(m\geq 7\) (if \(m\leq 6\) they must be constant). This paper is the continuation of a work with the same scope for functions [\textit{K. Kunikawa} and \textit{Y. Sakurai}, J. Geom. Anal. 31, No. 12, 11899--11930 (2021; Zbl 1486.53104)] and aims at a generalisation to metric flows of \textit{S.-t. Yau}'s Liouville theorem on complete manifolds [Commun. Pure Appl. Math. 28, 201--228 (1975; Zbl 0291.31002)]. The method is again to obtain local gradient estimates and nothing is said about the existence of such flows.