Convergence of the fractional Yamabe flow for arbitrary initial energy

Technical Summary: Convergence of the Fractional Yamabe Flow for Arbitrary Initial Energy

Problem Statement
The paper addresses the convergence of the fractional Yamabe flow, a geometric evolution equation defined on a closed manifold $M$ which is the conformal infinity of a Poincaré-Einstein manifold $(X, g_+)$. The flow seeks to evolve a metric $g(t)$ within a given conformal class to one with constant fractional scalar curvature $R^g_\gamma$. While the classical Yamabe flow (corresponding to $\gamma=1$) was proven to converge for arbitrary initial energy by Brendle (2005), the fractional case ($\gamma \in (0,1)$) had previously only been established under a restrictive "small initial energy" assumption (Chan, Sire, and Sun, 2021). The primary objective of this work is to remove this energy constraint and prove strong convergence for arbitrary initial energy, provided the fractional Positive Mass Conjecture holds.

Methodology
The authors adapt the strategy pioneered by Brendle for the local case to the non-local fractional setting. The core difficulty lies in the non-locality of the conformal fractional Laplacian $P^g_\gamma$, which prevents the direct application of pointwise product rules used in local PDE analysis. To overcome this, the authors employ the following techniques:

1. Extension Problem: They utilize the Caffarelli-Silvestre type extension problem, lifting the problem to a higher-dimensional manifold $(X, g_+)$ where the fractional operator is realized as a Dirichlet-to-Neumann map.
2. Blow-up Analysis: The proof relies on a detailed blow-up analysis of the sequence of solutions $u_\nu$ as time $t \to \infty$. The authors decompose the solution into a "regular" part $u_\infty$ and a sum of "bubbles" (concentrating solutions) centered at points $x_{k,\nu}$ with scales $\epsilon_{k,\nu}$.
3. Handling Non-local Interactions: A significant portion of the technical work involves deriving precise estimates for the interaction between these bubbles and the background geometry. Specifically, the authors compare "glue-and-extend" and "extend-and-glue" versions of Schoen's bubble (constructed by attaching Green's functions to standard fractional bubbles). They derive novel pointwise gradient estimates and asymptotics for the Poisson kernel of fractional conformal Laplacians, building on recent work by Mayer and Ndiaye.
4. Orthogonality and Projection: In the case where the limit function $u_\infty$ is non-zero, the authors construct finite-dimensional perturbations $\bar{u}_z$ of $u_\infty$ satisfying specific orthogonality conditions with respect to the kernel of the linearized operator. This allows them to control the energy functional via a projection operator $\Pi$.

Key Contributions and Results
The paper establishes the following main theorem:

• Theorem 1.2: Let $n \ge 2$ and $\gamma \in (0,1)$. Assuming the fractional Yamabe energy $Y_\gamma(M, [g_0]) > 0$, the first eigenvalue condition $\lambda_1(g_+) \ge n^2/2 - \gamma^2$, the vanishing of the mean curvature $H=0$ (if $\gamma \in (1/2, 1)$), and the validity of the Positive Mass Conjecture (with $A_0 > 0$), the fractional Yamabe flow converges strongly for arbitrary initial energy.

The proof proceeds by establishing a key estimate (Proposition 2.1) which bounds the difference between the current average curvature $s_\gamma(t_\nu)$ and the limit $s_\infty$ by a higher-order term of the $L^p$ norm of the curvature deviation. This estimate is derived by analyzing two distinct cases:

1. Vanishing Limit ($u_\infty \equiv 0$): The solution concentrates entirely into bubbles. The authors show that the energy of the solution is bounded by the sum of the energies of the bubbles, leading to a contradiction unless the number of bubbles is zero or the flow converges.
2. Non-vanishing Limit ($u_\infty > 0$): The solution converges to a non-trivial solution plus bubbles. The authors utilize the spectral properties of the linearized operator around $u_\infty$ and the constructed perturbations $\bar{u}_z$ to derive energy decay estimates.

A crucial technical contribution is the derivation of a novel pointwise gradient estimate (Equation 3.10) and the rigorous handling of the interaction terms between bubbles in the non-local setting, which were previously obstacles to extending Brendle's argument.

Significance and Claims
The authors claim that this work resolves the expected convergence theory for the fractional Yamabe flow, mirroring the classical result by Brendle but adapted to the non-local fractional setting. The paper explicitly states that the removal of the small energy assumption is the primary advancement.

The significance of the result is conditional: the convergence is proven "whenever the (fractional) positive mass conjecture is valid." The authors acknowledge that the Positive Mass Conjecture for fractional operators remains an open problem due to non-locality and a lack of tools, but they proceed by assuming its validity to demonstrate the convergence mechanism.

The paper does not claim to prove the Positive Mass Conjecture itself, nor does it propose new experimental applications. Instead, it provides a rigorous analytical framework showing that if the geometric mass condition holds, the flow behaves well regardless of the initial energy level. The authors note that while the convergence is stated in a Hölder sense in the main theorem, standard parabolic regularity theory suggests the convergence is actually smooth ($C^\infty$).