MATLAB Answers

how to reduce the running time (preallocate)

3 views (last 30 days)
Jisu Yang
Jisu Yang on 14 Jun 2020
Edited: Gaurav Garg on 17 Jun 2020
function[A,Y_linear2,ksi] = iteration(x0,y0,xp,yp,X_target,X_true,x,y,ksi0,Input);
syms omega phi kappa;
R_omega=[1 0 0 ; 0 cos(omega) -sin(omega) ; 0 sin(omega) cos(omega)];
R_phi=[cos(phi) 0 sin(phi) ; 0 1 0 ; -sin(phi) 0 cos(phi)];
R_kappa=[cos(kappa) -sin(kappa) 0 ; sin(kappa) cos(kappa) 0 ; 0 0 1];
R=R_omega*R_phi*R_kappa;
r11=R(1); r21=R(2);r31=R(3);r12=R(4);r22=R(5);r32=R(6);r13=R(7);r23=R(8);r33=R(9);
syms X0 Y0 Z0
for i = 1:length(X_target)
Y_linear2(2*i-1) = [x(i)-x0(i)];
Y_linear2(2*i)= [y(i)-y0(i)];
end
if Input == 'DSC00370.jpg'
A = [jacobian([x0(1)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([y0(1)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([x0(2)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([y0(2)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([x0(3)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([y0(3)-xp],[X0,Y0,Z0,omega,phi,kappa]);...
jacobian([x0(4)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([y0(4)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([x0(5)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([y0(5)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([x0(6)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([y0(6)-xp],[X0,Y0,Z0,omega,phi,kappa])];
elseif Input == 'DSC00371.jpg'
A = [jacobian([x0(1)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([y0(1)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([x0(2)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([y0(2)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([x0(3)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([y0(3)-xp],[X0,Y0,Z0,omega,phi,kappa]);...
jacobian([x0(4)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([y0(4)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([x0(5)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([y0(5)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([x0(6)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([y0(6)-xp],[X0,Y0,Z0,omega,phi,kappa]);...
jacobian([x0(7)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([y0(7)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([x0(8)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([y0(8)-xp],[X0,Y0,Z0,omega,phi,kappa])];
end
%%del_ksi = inv(A'*eye(6)*A)*A'*eye(6)*(Y_linear)';
n=1;
while (n<3);
del_ksi = (A'*eye(2*length(X_target))*A)\eye(6)*A'*eye(2*length(X_target))*(Y_linear2)';
n=n+1;
ksi0 = ksi0 + del_ksi;
x0(i) = xp-c*Nomx(X_true(i), ksi0(1),Y_true(i),ksi0(2),Z_true(i),ksi0(3))/Den(X_true(i), ksi0(1),Y_true(i),ksi0(2),Z_true(i),ksi0(3));
y0(i) = yp-c*Nomy(X_true(i), ksi0(1),Y_true(i),ksi0(2),Z_true(i),ksi0(3))/Den(X_true(i), ksi0(1),Y_true(i),ksi0(2),Z_true(i),ksi0(3));
if Input == 'DSC00370.jpg'
A = [jacobian([x0(1)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([y0(1)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([x0(2)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([y0(2)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([x0(3)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([y0(3)-xp],[X0,Y0,Z0,omega,phi,kappa]);...
jacobian([x0(4)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([y0(4)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([x0(5)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([y0(5)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([x0(6)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([y0(6)-xp],[X0,Y0,Z0,omega,phi,kappa])];
elseif Input == 'DSC00371.jpg'
A = [jacobian([x0(1)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([y0(1)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([x0(2)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([y0(2)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([x0(3)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([y0(3)-xp],[X0,Y0,Z0,omega,phi,kappa]);...
jacobian([x0(4)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([y0(4)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([x0(5)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([y0(5)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([x0(6)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([y0(6)-xp],[X0,Y0,Z0,omega,phi,kappa]);...
jacobian([x0(7)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([y0(7)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([x0(8)-xp],[X0,Y0,Z0,omega,phi,kappa]);jacobian([y0(8)-xp],[X0,Y0,Z0,omega,phi,kappa])];
end
for i = 1:length(X_target)
Y_linear2(2*i-1) = [x(i)-x0(i)];
Y_linear2(2*i)= [y(i)-y0(i)];
end
end
ksi
end
This is a part of my code. As i guess, it take so long because some variables keep changing during the while iteration.
In matlab, it shows an error message that certain variables are changing during the loop so that preallocation would help reducing the running time.

  1 Comment

dpb
dpb on 14 Jun 2020
"This is a part of my code. ..."
n=1;
while (n<3);
del_ksi = (A'*eye(2*length(X_target))*A)\eye(6)*A'*eye(2*length(X_target))*(Y_linear2)';
n=n+1;
ksi0 = ksi0 + del_ksi;
x0(i) = xp-c*Nomx(X_true(i), ksi0(1),Y_true(i),ksi0(2),Z_true(i), ...
And not enough part, either..."i" is undefined in the above segment which appears to probably be where the problems start.
Just start with the editor messages and work your way through -- what is the range of i for the above code? You've bounded a variable n, but it's never defined nor incremented to stop the otherwise ifinite loop the while has created.
Definitely nothing anybody can help with with only this piece to look at.

Sign in to comment.

Answers (1)

Gaurav Garg
Gaurav Garg on 17 Jun 2020
Edited: Gaurav Garg on 17 Jun 2020
Hi,
What could be a problem here is -
You are assigning values to arrays which haven't been allocated any space beforehand.
For an example -
You may try to use this code
a = ones(1,10)
for i = 1:10
a(i) = i;
end
rather than this one
for i = 1:10
a(i) = i;
end
Some of the variables where you can pre-allocate space are -
  1. y0
  2. x0
  3. Y_linear2
Note:
Probably, you are facing a warning and not an error. Your program would run perfectly fine even if you don't make the changes suggested.

  0 Comments

Sign in to comment.


Translated by