Hi, I am not sure what wrong with this code? Can I directly use mad function in CVX?

Target = 1.06;

delta = 0.3;

for k = 1:length(Target)

cvx_begin quiet

variables t z(10) x(4);

minimize(prob*(z+t)+ delta*mad(z+t));

subject to

z + t + B * ret *x >= L;
z + t >= 0;
z >= 0;
mu*x >= Target;

ones(1,4)*x == 1;

x >= 0;

cvx_end

end

erorr:

z = x.id_ == y.id_;

if isequal(x,[]), y = nan(‘like’,x); return; end

minimize(prob*(z+t)+ delta*mad(z+t));