MATLAB: Feasibility lsqlin vs. fmincon

fminconlsqlin

Hello everybody,
I have an objective function to minimize:
fun=@(x) sqrt(sum((x-d).^2)).
Instead of this function we can also minimize the function
fun=@(x) sum((x-d).^2) ,
so that we have a linear least squares problem. Additional, there are linear constraints. That's why I decided to use lsqlin:
 Eye = eye(size(d,1))
 [c,fval,residual,exitflag,output,lambda] = lsqlin(Eye,d,A,b,[],[],lb,ub).
For some inputs I get the message
    'Exiting: the constraints are overly stringent;
 no feasible starting point found.'
That's why I tried it with fmincon:
 f=zeros(length,1);
 x0=linprog(f,A,b,[],[],lb,ub);
 [c,fval]=fmincon(fun,x0,A,b,[],[],lb,ub);
Now the (for me) surprising result: fmincon provides the message 'Local minimum found that satisfies the constraints.' for the same inputs. So: Why does it work for fmincon whereas lsqlin doesn't work for the same inputs?
Thank you for your help!
Diana

Best Answer

You also get no solution from lsqlin if you supply an initial guess for x0 (like in the case where you use fmincon) ?
Best wishes
Torsten.
Related Question