Prev: Error when using xlswrite
Next: How to use ColorOrder for a added plot line?
From: jy on 20 Jul 2010 20:47
hello.
i optimized with fmincon some b-splines.
control points of b-splines are variations of my optimization
after some iterations, the control points chage to NAN.
i don't know the reason!! the knots of control points seem to be real and positive.
Knots have no problem.

I attach part of my object function
function [Heatlossrate] = obj_1axis(x)
global nDOF q_i q_f nMidCP g_iteration nDegree
global Pp2 Vt bulkmodulus Ahp_k Arp_k M_k
global maxQL maxSysRel g_tf g_gravity Vpos Vneg

tf=g_tf;
cp=zeros(nMidCP+4,nDOF);
timestep=tf/g_iteration;
for i=1:1:nDOF
cp(1,i)=q_i(i);
cp(2,i)=q_i(i);
for j=1:1:nMidCP
cp(2+j,i)= x(j+nMidCP*(i-1)); % q_f(i) q_f(i)]';
end
cp(nMidCP+3,i)=q_f(i);
cp(nMidCP+4,i)=q_f(i);
end
n_number=4+nMidCP-1; k=nDegree+1;
knots=make_knots(n_number,k);
knots= knots/knots(1,n_number+k+1)*tf;
From: us on 20 Jul 2010 21:40
"jy " <eule3029(a)naver.com> wrote in message <i25g27$m9b$1(a)fred.mathworks.com>...
> hello.
> i optimized with fmincon some b-splines.
> control points of b-splines are variations of my optimization
> after some iterations, the control points chage to NAN.
> i don't know the reason!! the knots of control points seem to be real and positive.
> Knots have no problem.
>
> I attach part of my object function
> function [Heatlossrate] = obj_1axis(x)
> global nDOF q_i q_f nMidCP g_iteration nDegree
> global Pp2 Vt bulkmodulus Ahp_k Arp_k M_k
> global maxQL maxSysRel g_tf g_gravity Vpos Vneg
>
> tf=g_tf;
> cp=zeros(nMidCP+4,nDOF);
> timestep=tf/g_iteration;
> for i=1:1:nDOF
> cp(1,i)=q_i(i);
> cp(2,i)=q_i(i);
> for j=1:1:nMidCP
> cp(2+j,i)= x(j+nMidCP*(i-1)); % q_f(i) q_f(i)]';
> end
> cp(nMidCP+3,i)=q_f(i);
> cp(nMidCP+4,i)=q_f(i);
> end
> n_number=4+nMidCP-1; k=nDegree+1;
> knots=make_knots(n_number,k);
> knots= knots/knots(1,n_number+k+1)*tf;

why do you stupidly repeat your OP after ~5 minutes(?)...

us
 | 
Pages: 1
Prev: Error when using xlswrite
Next: How to use ColorOrder for a added plot line?