First, the optimization variable is X, defining x=X*X’, matrix B, C1 and C2.Next,my code is as follows:

%----------------Parameter---------------

```
ZT = [13516 0.079 0.079 0.079;...
0.079 13516 0.0152 0.0152;...
0.079 0.0152 13516 0.0152;...
0.079 0.0152 0.0152 13516];
ZR = [6402.3 0.0179;...
0.0179 6402.3];
M = [-0.2527 0.0105 0.0046 0.0066;...
-0.2527 0.0046 0.0105 0.0066];
Rm1 = [56 0;0 0];
Rm2 = [0 0;0 56];
w = 6.78*2*pi*10^6;
B = ZT+(w^2)*M.'*(ZR^-1)*M;
C1 = ((ZR^-1)*M)'*Rm1*(ZR^-1)*M;
C2 = ((ZR^-1)*M)'*Rm2*(ZR^-1)*M;
beta = linspace(0,80,9);
n = 4;
num = 2;
%----------------CVX-------------------
for k = 1:length(beta)
cvx_begin sdp
variables x(n)
X = x*x';
minimize( 0.5*trace(B*X) )
subject to
X == semidefinite(n);
trace(C1*X)>=(2*beta(k))/w^2;
trace(C2*X)>=(2*beta(k))/w^2;
cvx_end
obj1(k) = cvx_optval;
obj2(k) = (num*beta(k))/cvx_optval;
end
```

so CVX does not support x*x’,error reporting.Is there any good way to solve it?