The non convex energy for a ribbon has the form:
The relaxed energy is is the largest convex minorant of . For this energy, an explicit formula is contained in the paper with RP. A key point in that formula is that the determinant for a 2x2 matrix is quadratic, and hence it nicely interacts with the quadratic energy.
The complementary energy is .
We can gather the components of into a vector . We can also introduce the matrix
Also, we can set
Then we have
Then we have where
Note that since if , we can also write
We note that the above maximization problem needs not have a unique solution. Let be any such solution. Then,
To solve the variational problem we introduce the Lagrangian . The optimality conditions are
In particular, taking the scalar product of both sides of the first equation with , we obtain