Stochastic Willmore Flow in Planar Curves

• Stochastic Willmore flow is a stochastic generalization of planar curve evolution that incorporates Willmore energy minimization and Brownian forcing.
• It reformulates the evolution as a one-phase Stefan problem and a quasilinear SPDE, enabling the use of maximal Lp-regularity methods.
• The approach guarantees local existence, uniqueness, and blow-up alternatives, offering precise analytical insights into curvature singularities.

The stochastic Willmore flow is a stochastic generalization of the classical Willmore (free elastic) curve flow in the plane, designed to model the evolution of planar curves under the combined influence of Willmore energy minimization and random (Brownian) forcing. By equivalently recasting this stochastic geometric evolution as a one-phase Stefan-type free boundary problem for the curvature, this subject connects geometric analysis, stochastic partial differential equations (SPDEs), and the maximal regularity theory for quasilinear evolution equations (Yan, 25 Nov 2025).

1. Stochastic Willmore Flow: Geometric and Analytic Formulation

Let $\gamma_t$ denote a family of immersed planar curves of time-dependent length $L(t)$, parameterized by arc-length $s \in [0, L(t)]$. The curvature at $(s, t)$ is written $k(s, t)$. The stochastic Willmore flow describes the dynamics where the normal velocity $V(s, t)$ of each point on the curve is given by: $$V(s, t) = -\left(\partial_{ss}k + \tfrac{1}{2}k^3\right) - \circ \frac{dW_t}{dt}$$ Here, $W_t$ is a cylindrical Wiener process acting in the normal direction, and $\circ$ indicates the use of Stratonovich integration.

The corresponding stochastic free boundary problem for the curvature and length is: $$\begin{cases} dk(s, t) = \Bigl[ -\partial_{ss}(\partial_{ss}k + \tfrac{1}{2}k^3) - k^2 (\partial_{ss}k + \tfrac{1}{2}k^3) \Bigr]\,dt + \left(\partial_{ss} + k^2\right) \circ dW_t, & s \in [0, L(t)],\, t>0 \ dL(t) = \displaystyle\int_0^{L(t)} k(\partial_{ss}k + \tfrac{1}{2}k^3)\,ds\,dt - \int_0^{L(t)} k\,\circ dW_t, & t>0 \ k(s,0) = k_0(s), \quad L(0) = L_0 \end{cases}$$ The $L^2$-gradient structure of the drift and the form of the stochastic perturbation reflect the variational origin of the deterministic Willmore flow.

2. Arc-Length Parameterization, Intrinsic Quantities, and Normal Flow Law

Given a general parameterization $\gamma(u, t)$, arc-length $s(u, t)$ and total length $L(t)$ are defined by: $$s(u, t) = \int_0^u |\partial_r \gamma(r, t)|\,dr, \qquad L(t) = \int_0^{2\pi} |\partial_u \gamma(u, t)|\,du$$ The tangent and normal vectors, as well as the tangential angle $\theta(s, t)$, satisfy: $$T = \partial_s \gamma = (\cos \theta, \sin \theta), \qquad \partial_s T = -kN, \qquad \partial_s \theta = k$$ The intrinsic (arc-length based) evolution for curvature and length under a normal velocity $V$ yields: $$\partial_t k = \partial_{ss} V + k^2 V, \qquad \partial_t L = -\int_0^{L(t)} k V\,ds$$ Setting $V$ as above, this system matches the Stefan-type SPDE structure.

3. Equivalence to a Stochastic One-Phase Stefan Problem

Interpreting the stochastic Willmore flow as a one-phase Stefan problem: $$\begin{cases} \partial_t k = \partial_{ss} V + k^2 V, & s \in [0, L(t)] \ \partial_t L(t) = -\int_0^{L(t)} k(s, t) V(s, t)\,ds \end{cases}$$ with $$V = -(\partial_{ss}k + \tfrac{1}{2}k^3 + \circ \dot{W})$$ encodes the free boundary evolution within the stochastic problem. The “Stefan condition" asserts that the free endpoint moves with the normal velocity determined by integrating $kV$.

4. Quasilinear SPDE Reformulation in a Banach Space Setting

Introduce the rescaled variable $r = s/L(t) \in \mathbb{T} = \mathbb{R}/\mathbb{Z}$ and function $f(r, t) = k(r L(t), t)$. The pair $X = (f, L)$ satisfies: $$dX_t + A(X_t)\,dt = F(X_t)\,dt + G(X_t)\,dW_t, \qquad X \in X_1 \subset X$$ with

• $X = L^q(\mathbb{T}) \times \mathbb{R}$, $X_1 = W^{4,q}(\mathbb{T}) \times \mathbb{R}$;
• Principal operator: $$A(f, L) = \begin{pmatrix} L^{-4} \partial_{rrrr} + \left(\tfrac{5}{2}L^{-2}f^2 - 2\pi^2r^2 L^{-2}\right) \partial_{rr} & 0 \ 0 & 0 \end{pmatrix}$$ which is uniformly parabolic in $r$, with $H^\infty$-calculus and $R$-bound properties verified in weighted Sobolev spaces;
• The drift $F(f, L)$ and noise $G(f, L) \in \gamma(\mathbb{R}; X^{1/2})$ are explicit algebraic functions of $f$, its derivatives, and various moments.

This reformulation enables the deployment of modern maximal $L^p$-regularity for quasilinear stochastic evolution equations in UMD Banach settings.

5. Existence, Uniqueness, and the Blow-Up Alternative

With $1 < p, q < \infty$ and the condition

$\frac{4}{p} + \frac{1}{q} < 1,$

for any initial $$(k_0, L_0) \in B^{4-4/p}_{q, p}(\mathbb{T}) \times (0,\infty)$$ there exists a unique $L^p$-maximal local strong solution $(k, L, \sigma)$, with stopping times $\sigma_n \uparrow \sigma$, such that for all $n$: $$(k, L) \in L^p\big(\Omega; H^{\theta, p}([0, \sigma_n]; H^{4(1 - \theta), q}(\mathbb{T}) \times \mathbb{R})\big) \cap L^p\big(\Omega; C([0, \sigma_n]; B^{4 - 4/p}_{q, p}(\mathbb{T}) \times \mathbb{R})\big)$$ for all $\theta \in [0, 1/2)$, and $\sigma > 0$ almost surely.

The maximal solution is characterized by the following blow-up alternative: $$\P\left\{ \sigma < T,\; \|k\|_{L^p(0, \sigma; W^{4, q})} < \infty,\; 0 < L(t) < \infty,\; (k, L) \text{ uniformly continuous on } [0, \sigma) \right\} = 0$$ That is, either the solution is global, or

$$\limsup_{t \to \sigma} \left( \|k(t)\|_{W^{4, q}} + L(t)^{-1} \right) = \infty$$

meaning singularities in curvature or the collapse of the curve may occur.

6. Geometric Singularities and Regularization

Blow-up is restricted to two geometric scenarios: either $\|k(\cdot, t)\|_{W^{4, q}} \to \infty$ (appearance of curvature singularities) or $L(t) \to 0$ (curve collapses to a point) as $t \uparrow \sigma$.

Further features include:

• For closed curves, the topological invariant $\int_0^{L(t)} k\,ds = 2\pi$ is preserved, used to eliminate boundary noise.
• The deterministic Willmore flow is the $L^2$-gradient flow of the bending energy $\frac{1}{2} \int k^2\,ds$; in the stochastic case, martingale terms appear in the energy balance.
• Instantaneous regularization: for $t > 0$, the curve is almost surely $C^{3, \alpha}$ in space.

7. Significance and Connections

The stochastic Willmore flow presents, for the first time, a stochastic one-phase Stefan formulation for planar Willmore flow. Its conversion to a fully intrinsic quasilinear SPDE and the application of stochastic maximal $L^p$-regularity theory enable rigorous local-in-time existence, uniqueness, and analytic characterization of singularities for randomly forced geometric curve evolutions. These results demonstrate the interplay between geometric analysis, stochastic analysis, and quasilinear maximal regularity in an evolving planar geometry context (Yan, 25 Nov 2025).