Minimum Surface AreaSOCP > SOC inequalities | Standard Forms | Group sparsity | Applications > Minimum Surface Area | Next
The minimum surface area problemFunctional problem statementConsider a surface in that is described by a function from the square to . The corresponding surface area is The minimum surface area problem is to find the function which minimizes the area , subject to boundary values. To be specific, we will assume that we are given values of on the left and right side of the square, that is where and are two given functions. The above is an infinite-dimensional problem, in the sense that the variable is a function, not a finite-dimensional vector. DiscretizationWe can discretize the square with a square grid, with points , , where is an integer, and where the (uniform) spacing of the grid. We represent the variable of our problem, , as a matrix , with elements . Similarly, we represent the boundary conditions as vectors of length , , To approximate the gradient, we start from the first-order expansion of a function of two variables, valid for some small increment : We obtain that the gradient of at a grid point can be approximated as SOCP formulationThe discretized version of our problem is thus The CVX syntax for this problem can be as follows. CVX syntax
>> % input: left_vals and righ_vals, two row vectors of length K+1 >> h = 1/K; cvx_begin variables F(K+1,K+1) variables T(K,K) minimize( sum(T(:)) ) subject to for j = 1:K, for i = 1:K, norm([K*(F(i+1,j)-F(i,j)); K*(F(i,j+1)-F(i,j)); 1],2) <= T(i,j); end, end F(1,:) == left_vals; F(K+1,:) == right_vals; cvx_end Examples
It is interesting to compare the minimal surface area with one that is obtained by squaring the norms. This corresponds to the QP |