if my convex optimization variable is a matrix X,Y can one of the constraints be trace(x^2)>trace(y^2) and can there be another constraint bounding the determinant of the matrix X, smth like det(X)<=5.
I was curious if these operations were convex and implementable?
Thanks for ur reply
No, neither one of those is convex. You can find a complete list of the functions supported by CVX in the documentation.
where can i find an exact list of all functions supported by cvx, or is it just based on whether the function is convex or not?
The user’s guide has a complete list. And it’s not just a matter of limiting yourself to the accepted function list; you must combine those functions according to the rules, again given in the user’s guide. I suspect, however, that your model is not compatible with CVX, given the examples here.