Hello, I have a quasi-concave problem that can be solved by the bisection algorithm, in each step solving a sequence of convex feasibility problem. I need to run it for hundred times however in the 91st iteration, it has errors, Undefined function or variable ‘cvx_check_dimlist’. I attached the code, concave problem and errors below. Actually when I just used the 90 values I can obtained the result which I want to.

% Max-min power control
R_dk_min = min(R_dk);
t_min = 2^R_dk_min-1;
t_max = 2.5;
epsi = max(t_min/5,0.01);
v_k1_phi = zeros(K,K);
I_k = abs(eye(num_UEs)-1);
for k=1:K
for k_dash=1:K
v_k1_phi(k_dash,k)=phi_k(:,k_dash)’*phi_k(:,k);
end
end
v_k1_phi = v_k1_phi.*I_k;

while(t_max-t_min>epsi)
t=(t_max+t_min)/2;
cvx_begin sdp quiet
variables var_sigma(M,K) var_rho(K,K) var_theta(M,1)
minimize(0)
subject to
for k=1:K
norm([v_k1_phi(:,k).*var_rho(:,k);sqrt(beta_mk(:,k)).*var_theta;1/sqrt(rho_d)]) <= 1/sqrt(t)*gamma_mk(:,k)’*var_sigma(:,k);
end
for m=1:M
norm(sqrt(gamma_mk(m,:)).*var_sigma(m,:)) <= var_theta(m);
var_theta(m) <= 1;
var_theta(m) >= 0;
end
for k=1:K
for k_dash=1:K
sum(gamma_mk(:,k_dash).*beta_mk(:,k)./beta_mk(:,k_dash).*var_sigma(:,k_dash)) <= var_rho(k_dash,k);
end
end
for m=1:M
for k=1:K
var_sigma(m,k) >= 0;
end
end
cvx_end
% Feasible bisection
if contains(cvx_status,‘Solved’) % feasible
t_min = t;
else % not feasible
t_max = t;
end
toc
end
R_dk_opt_min(round)=log2(1+t_min);

You have not provided input data, so it is not reproducible. I suggest you use the MATLAB diary and remove the quiet option Then check in detail the output of the 90th iteration and input to 91st iteration, and see exactly what is going on. That said, I don;t know why that error message would occur.