problem by using symengine

Hallo,
i am trying to program this code but when i try to use jacobian by A matrix i became wrone msg that error by using symengine
but it works with B matrix !!
the msg also says "unable to convert experssion into double array"
%aufgabestellung3
clear all;clc;close all;
L=[0 1 4 9 0 1 2 3]';[n,~]=size(L);
r=4; n=8; u=2;
Lsym=sym('Lsym',[n,1]);
v0=zeros(n,1); vsym=sym('vsym',[n,1]);
X0=[0 0]';
[u,~]=size(X0);
Xsym=sym('Xsym',[u,1]);
r=4;
%sto Modell
Qll=eye(n);
P=inv(Qll);
%funk Modell % Ansatz: L=[x;y]; 0=ax+b-y; X=[a b];
% A=[x ones(r,1)]; B=[X0(1,1)*eye(r) -1*eye(r)];
for i =1:r
f(i,1)=Xsym(2,1)+Xsym(1,1)*(Lsym(r+i,1)+vsym(r+i,1))-(Lsym(i,1)+vsym(i,1));
end
it=0;abr=1;
while it<10 & abr>10^-20
B=double(subs(jacobian(f,vsym),[vsym;Xsym],[v0;X0]));
% A=[x+v0(1:r,1) ones(r,1)]
A=double(subs(jacobian(f,Xsym),[vsym;Xsym],[v0;X0])); % here i become the wrong !
w0=double(subs(f,[Lsym;Xsym],[L;X0]));
w=-B*v0+w0
% Loesungsalgorithmus (Niemeier S. 178 Gl. 5.2.30)
iBQB=inv(B*inv(P)*B'); % Substitution
Q22=-inv(A'*iBQB*A);
Q12=-iBQB*A*Q22;
Q21=Q12';
Q11=iBQB*(eye(r)-A*Q21);
% Gleichung 5.2.31
k=-Q11*w;
xd=-Q21*w;
Xd=X0+xd
Qll=inv(P);
v=Qll*B'*k;
Ld=L+v;
wd=zeros(r,1)
wd=double(subs(f,[vsym;Xsym],[v;Xd]));
abr=max(abs(wd))
it=it+1
X0=Xd
end

2 个评论

Your code is assuming that the jacobian of f with respect to Xsym gets rid of all Lsym variables, but it does not. You have sub-expressions that are Lsym times something involving Xsym variables, so the Lsym remain in the jacobian.
i edited it i have to insert L as a vector not as sympolic variable , i have no idea why
but the Prof at the uni didnt use jacobian for the Partial differentiation he inserted the both matrix A,B direct
A=[x+v0(1:r,1) ones(r,1)]
B=[X0(1,1)*eye(r) -1*eye(r)]
but i thought by jacobian or using diff its faster and safer
but there is a problem by Partial differentiation
by me B at the final is
B =
0 0 0 0 -1 0 0 0
0 0 0 0 0 -1 0 0
0 0 0 0 0 0 -1 0
0 0 0 0 0 0 0 -1
by prof B =
3.2418 0 0 0 -1.0000 0 0 0
0 3.2418 0 0 0 -1.0000 0 0
0 0 3.2418 0 0 0 -1.0000 0
0 0 0 3.2418 0 0 0 -1.0000
A =
0 1
1 1
2 1
3 1
A =
0.3838 1.0000
0.7524 1.0000
1.6843 1.0000
3.1795 1.0000
it means there is no correction happened

请先登录,再进行评论。

回答(0 个)

类别

帮助中心File Exchange 中查找有关 MATLAB 的更多信息

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by