CVX warning: Too many output arguments

Nonconvex
1

I was trying to find matrix V optimal:

n = 51;

cvx_begin
variable V(n,n) hermitian toeplitz
maximize ( abs(trace(R * V)) )
subject to
diag(V) == 1
V == hermitian_semidefinite( n )
cvx_end

where R is a n*n complex matrix generated before.

After running the program, I received error message saying that:
“Error using type”
“Too many output arguments”
“Error in maximize( abs(trace(R * V)) )”

But when I change the “maximum” to “minimum”, I can get an optimal value of V.

Can someone help me solve this problem?

2

abs(trace(R * V)) is a convex expression which can be minimized, not maximized. I don’t know why the error (warning?) message is so cryptic.