```
indent preformatted text by 4 spaces
Nt = 10;
Ne = 4;
sigma = 1;
H = rand(Nt,Ne);
Pmax = 1;
cvx_begin sdp
variable F1(Nt,Nt) hermitian semidefinite
variable F2(Nt,Nt) hermitian semidefinite
minimize (lambda_max((H'*F2*H+sigma^2*eye(Ne))\(H'*F1*H)))
subject to
real(trace(F1) + trace(F2))<=Pmax
cvx_end
```

You can’t use this formulation in CVX.

To model generalized eigenvalue problems, read Linear Matrix Inequalities in System and Control Theory - Stephen Boyd, Laurent El Ghaoui, Eric Feron, and Venkataramanan Balakrishnan. Then form a DCP based on the approaches there.

And please don’t keep starting multiple threads with the same topic, which makes clean up work for me.

thank you very much, i got it. thank you