Complex sparse LDL is not yet available

1

close all
clear all
clc

load data1.mat

zeta =0.001;

cvx_begin sdp quiet
variable W(K,I)
variables t1 t2
expression x(Mt,K)
expression X(Mt,L)
expression R(Mt,L)
expression delta_x(K-1,K-1)
expression t
expression delta_x_k
expression Pt
expression PJ
expression Rj
minimize t

A=p_mk+Wp_aj;
x(:,1:K )=a1(:,1:K )+a2(W(1:K ,:).’);

for k=1:size(x,1)
x_k=x(k,:);
for i=1:size(x,2)
for j=i:(size(x,2)-1)
delta_x(i,j)=abs(x_k(:,i)-x_k(:,j+1));
end
end
delta_x_k(k)=max(delta_x(:));
end

sum_1=sum(abs((A(:))));
sum_1<=t1;
t1<=1.5*E;

sum_2=sum(delta_x_k(:));
sum_2<=t2;
t2<=1.5*E;
t=t1+t2;
W(:)>=0;

X=zeros(Mt,1);
for k=1:K
X=[X,kron(ones(1,c(k)),x(:,k))];
end
X=X(:,2:end);
vX=X(:);

quad_form(vX’,ASJ)<=MtLzeta;

cvx_end

2

Please omit some redundancy variables. And the message is

Error using ldl
Complex sparse LDL is not yet available.

Error in cvx/quad_form (line 185)
[ R, DD, prm ] = ldl( sparse( Q ), ‘upper’, ‘vector’ );

Error in cvx_problem_test (line 54)
quad_form(vX’,ASJ)<=MtLzeta;

ASJ is a Positive semidefinite Hermitian matrix.

3

At last, I divided the matrix compute into vector compute. Then the problem is resolved.