When I use cvx to solve SDP problem, how to avoid the solution result being 0 or sparse?

How to add constraint R>0, where R is a complex hermitian toeplitz matrix?

Thanks!

You need to provide more specific information about your problem.

What do you mean by `R > 0`

? is that a semidefinite constraint? if so, use

`R >= small_positive_number*eye(n)`

if in sdp mode, or

`R - small_positive_number*eye(n) == semidefinite(n)`

if not in sdp mode. small_positive_number is perhaps 1e-4 or 1e-5.

If i use R >= small_positive_number*eye(n), it will report error “Disciplined convex programming error:

Invalid constraint: {complex affine} >= {constant}”.

That is for sdp mode. Use `... == semidefinite`

… if not in sdp mode.

If you actually want elmentwise inequality, then you will have to use `real(...) >= 0`

and/or `imag(...) >= 0`

or whatever it is that you want.

But if I use … == semidefinite…, the elements of solution result will be 0 or sparse.

By that, I meant

`R - small_positive_number*eye(n) == semidefinite(n)`

Note I corrected a typo in my previous post where I had `== 0`

instead of ` == semidefinite(n)``

Or use the … >= 0 if using sdp mode.

Because I have no idea what your problem is, I don’t know whether this will resolve matters for you.