Min abs of a convex fun

(ciel bleu) #1

hello CVXers
A and B are my variables and have some lmi constraints.
Is there a way to minimize magnitude of norm(A,‘fro’)-sqrt(trace(B)) as my cost function using cvx?

Any help is greatly appreciated

(Mark L. Stone) #2

This is non-convex. Consider f(x) = abs(abs(x)-1), which equals 0 when evaluated at x = -1 and 1, and equals 1 when evaluated at their convex combination 0, thereby disproving convexity.

I commend you again to read, carefully, Why isn't CVX accepting my model? READ THIS FIRST! .

1 Like