Condition number of a matrix

Hi all,

I wanted to know if CVX has capabilities to handle the quasi-convex constraint of the condition number of a matrix, either as an inequality or as a objective function?

Perhaps, it can be recast as a problem that can be solved by CVX? Any thoughts or ideas would be helpful. Thanks!


The condition number of a symmetric positive definite matrix is indeed quasiconvex, so you can optimize over it by solving a sequence of convex optimization problems. I do not believe that the condition number is quasiconvex in the indefinite or nonsymmetric cases, however. (In particular the minimum singular value of a matrix is not concave.)