Planar Wulff shape is unique equilibrium (Q4829903)

Let \(\phi\) be a continuous norm in \(\mathbb R^2\), i.e., \(\phi:\mathbb R^2\to\mathbb R^+\) is a nonnegative, convex, homogeneous function on \(\mathbb R^2\). Let \(c: S^1\to\mathbb R^2\) be a closed, connected, rectifiable curve parametrized by the arc length \(s\), and set \[ \psi(c)=\int_c \phi(c'(s))\,\, ds. \] We say that \(c=c_0\) is in equilibrium if for all smooth variations \(c_t(s)+f(s,t)\), preserving the area enclosed by \(c\), the derivative of \(\psi(c_t)\) is initially nonnegative. The \textit{Wulff shape} \(W_\phi\) of \(\phi\) is the boundary of the unit ball in the dual norm \(\phi^*\) of \(\phi\). It is known that, for a given \(\phi\), among all curves enclosing the same area, the boundary of the \(\phi^*\) ball \(B\) (the Wulff shape) minimizes \(\int_{\partial B} \phi(n)\, ds\); here \(n\) is the unit normal of \(c(s)\). The main result of this paper is that, for any continuous norm on \(\mathbb R^2\), an equilibrium, closed, connected curve \(c\) must be a (scaled) Wulff shape (possibly with integer multiplicity if nonembedded curves are allowed).