# Matrix dimension is incongruent mentioned in file cvxprob/eliminate row 137

I persume it is log_det(Omega_o) leads to this problem, where Omega_o is a matrix optimization variable
Anyone knows how to solve this problem, thank you bery much.
Code:
cvx_begin
variable R_Int_o(RRH_Num * RRH_Ant,RRH_Num * RRH_Ant) complex
variable R_User_o(RRH_Num * RRH_Ant,RRH_Num * RRH_Ant,UE_Num) complex
% variable R_Radar_o(RRH_Num * RRH_Ant, RRH_Num * RRH_Ant * RRH_Num * RRH_Ant) complex
variable alpha_o
variable Omega_o(RRH_NumRRH_Ant,RRH_NumRRH_Ant) complex
expression obj_sub1
obj = 0;
for grid = 1 : Tot_Grid
obj_sub1 = obj_sub1 + real( pow_abs(a_Steer(:,grid)’ * Theta_temp * Cas_IR_G * R_Int_o * Cas_IR_G’ * Theta_temp’ * a_Steer(:,grid) - alpha_o * Des_Beam(grid),2) );
end
minimize 1 / Tot_Grid * obj_sub1
subject to

``````            (R_Int_o - sum(R_User_o,3) - Omega_o)  == hermitian_semidefinite(RRH_Num * RRH_Ant)
R_Int_o == hermitian_semidefinite(RRH_Num * RRH_Ant)
for r = 1 : RRH_Num
real( trace(R_Int_o( (r-1)*RRH_Ant+1 : r*RRH_Ant , (r-1)*RRH_Ant+1 : r*RRH_Ant) ) ) == PR_mW
end
for ue = 1 : UE_Num
(1 + Tau^-1) * real( Int_Cha(ue,:) * R_User_o(:,:,ue) * Int_Cha(ue,:)' ) - real( Int_Cha(ue,:) * R_Int_o * Int_Cha(ue,:)' ) - n_sigma >= 0;
R_User_o(:,:,ue) == hermitian_semidefinite(RRH_Num * RRH_Ant);
end
for r = 1 : RRH_Num
real(log_det(Sigma(:,:,r))) - RRH_Ant ...
+ real(trace( Sigma(:,:,r) \ R_Int_o( (r-1)*RRH_Ant+1:r*RRH_Ant,(r-1)*RRH_Ant+1:r*RRH_Ant ) ))...
-real(log_det(E(:,:,r)' * Omega_o * E(:,:,r)))- C_l_ <= 10^-10;
end
Omega_o == hermitian_semidefinite(RRH_Num * RRH_Ant)
cvx_end``````

Try reinstalling CVX, in a new MATLAB session. Use CVX 2.2 If you can, try using Mosek as solver. Otherwise follow the instructions at CVXQUAD: How to use CVXQUAD's Pade Approximant instead of CVX's unreliable Successive Approximation for GP mode, log, exp, entr, rel_entr, kl_div, log_det, det_rootn, exponential cone. CVXQUAD's Quantum (Matrix) Entropy & Matrix Log related functions

Re-open MATLAB. see if it works. It would be nice if the Chinese inside were translated for the experts.

I try Mosek with cvx 2.2 and add a real function right inside the function log_det seems like it works, thank you very much