A speed preserving Hilbert gradient flow for generalized integral Menger curvature (Q6165634): Difference between revisions

Integral Menger curvature is one of several geometrically defined curvature functionals that are used in geometric knot theory to separate different knot classes by infinite energy barriers. It is defined as a triple integral evaluated on absolutely continuous closed curves \(\gamma\in\mathrm{AC}(\mathbb{R}/\mathbb{Z}, \mathbb{R}^n)\), namely \[ \mathcal{M}_p(\gamma) = \int\int\int_{(\mathbb{R}/\mathbb{Z})^3}\frac{|\gamma'(u_1)||\gamma'(u_2)||\gamma'(u_3)|}{R^p(\gamma(u_1), \gamma(u_2), \gamma(u_33))}du_1du_2du_3 \] For \(p>3\), this geometric curvature energy has regularizing properties. Furthermore, for \(p=3\) it ensures self avoidance. The combination of these effects can be used to minimize integral Menger curvature within any prescribed given tame knot class and to bound classic knot invariants and therefore the number of knot classes representable below given energy values. This renders \(\mathcal{M}_p\) as a valuable instrument in geometric knot theory. In the present paper, the authors establishes long-time existence for a projected Sobolev gradient flow of generalized integral Menger curvature in the Hilbert case and provide \(\mathcal{C}^{1,1}\)-bounds in time for the solution that only depend on the initial curve. The self-avoidance property of integral Menger curvature guarantees that the knot class of the initial curve is preserved under the flow, and the projection ensures that each curve along the flow is parametrized with the same speed as the initial configuration. Moreover, they show how to discretize and simulate this flow numerically, in a manner that has substantially higher efficiency than in the corresponding numerical \(L^2\) gradient descent or other optimization methods.