In an assignment A(I) = B, the number of elements in B and I must be the same.

at first i wrote below code to solve mechanism equations and find best parameters for my linkages

my first code:

l1=x(:,1);
l2=2;
l3=x(:,3);
l4=x(:,4);
l5=3;
l6=4;
theta1=0*pi/180;
alpha=acos((l3.^2+l5.^2-l6.^2)/(2*l3*l5));  % tha angle of coupler triangle

%%Gerashof principle %%

if l2+l3<=l1+l4;
    A=1;
else
    A=0;
end

R1=l1*exp(1i*theta1); %%length of 1 body

n=400; %%points
u=0; %%initial point for creating loop

for t2=0:2*pi/n:2*pi %%theta2 and step which it goes
    u=u+1;

      R2=l2*exp(1i*t2); %%length of 2 body
      D=R1-R2; %%length of interact of 1 and 2 bodies
      d=norm(D); %%mesure of interact body
      tD=angle(D); %%angle of interact body wich called "D"

      delta=abs(acos((l3.^2.+d.^2-l4.^2)/(2*l3*d))); %%angle between two bodies
      sigma=abs(acos((l4.^2.+d.^2-l3.^2)/(2*l4*d))); %%angle between two bodies

      theta3(u,1)=tD + delta;
      theta2(u,1)=t2 %%tehta 2
      P=A*(l2*exp(1i*t2)+l5*exp(1i*(alpha+theta3(u,1)))); %%length of distance between input and specific point
      px(u,1)=real(P); %%real measure
      py(u,1)=imag(P); %%imagine measure

But i encountered with this Error:

??? Error using ==> plus
Matrix dimensions must agree.

Error in ==> costFunction at 42
    theta3(u,1)=tD + delta;

then i changed that to this code:

l1=x(:,1);
l2=2;
l3=x(:,3);
l4=x(:,4);
theta1=0*pi/180;

%%Gerashof principle %%

if l2+l3<=l1+l4;
    A=1;
else
    A=0;
end

R1=l1*exp(1i*theta1); %%length of 1 body

n=3; %%points
u=0; %%initial point for creating loop

for t2=0:2*pi/n:2*pi %%theta2 and step which it goes
    u=u+1;

      R2=l2*exp(1i*t2); %%length of 2 body
      D=R1-R2; %%length of interact of 1 and 2 bodies
      d=norm(D); %%mesure of interact body
      t_D=angle(D) %%angle of interact body wich called "D"

      Delta=abs(acos((l3.^2+d.^2-l4.^2)/(2*l3*d))) %%angle between two bodies
      sigma=abs(acos((l4.^2+d.^2-l3.^2)/(2*l4*d))) %%angle between two bodies
  tD=t_D
  delta=Delta
  for nn=1:n
      theta3(nn)=tD+delta(nn)

end
    theta2(u,1)=t2 %%theta 2
%     theta3(u,1)=delta+tD %%theta3 

      P=A*(l2*exp(1i*t2)+l3*exp(1i*(theta3(u,1)))); %%length of distance between input and specific point

      px(u,1)=real(P); %%real measure
      py(u,1)=imag(P); %%imagine measure

    if true
      % code
    end

now i encounter with this Error:

???  In an assignment  A(I) = B, the number of elements in B and
 I must be the same.

Error in ==> costtFunction at 41
    theta3(nn)=tD+delta(nn)

i tried and read alot to solve this code but i couldn't.

would you please guide me ?

regards

[Image Analyst]

What is x? I couldn't even run your code without it.

[Morteza.Moslehi]

this is a function and i wanted to run it in coco algorithm to find best parameters for l1,l3,and l4.

[Image Analyst]

x is not a function. It can't be. I don't know why you just don't say what x is.

Anyway, I made up something for x, a 30 by 4 columns array. If x has N rows then tD is an N by 1 column vector and delta is an N by N square matrix. Thus you cannot add them because they are different shapes. You need to go over the logic some more.

And you shouldn't do this:

for t2=0:2*pi/n:2*pi

You should use integers and then get the angle inside the loop. Like

theta2 = t2=0:2*pi/n:2*pi 
for t2= 1 : length(theta2)
    thisTheta2 = theta2(t2);

[Niels]

thats exactly what i mean, use the command

keyboard

to check the size of td and delta and tell us

1.the size

2. if it has the size u wanted it to have

Answer 1

[Morteza.Moslehi]

Answer: i should write inputs same below:

cost=zeros(size(x,1),1);
for ii= 1:size(x)
l1=x(ii,1);
l2=2;
l3=x(ii,3);
l4=x(ii,4);
l5=3;
l6=4;

Answer 2

[Niels]

i tried to run it, but i could not at that moment i asked myself what your x might be

can u tell me?

if it is a vecor why would you type

l1=x(:,1);

if it is a matrix -> here is your problem

because then your R1 is also a vector, same for D and td

thats why u cant set theta3(1,1) a vector which is supposed to be a scalar

use

keyboard
to check if i am right

replace this part of your code and run it, afterwards type l1 in the cmd window

delta=abs(acos((l3.^2.+d.^2-l4.^2)/(2*l3*d))); %%angle between two bodies
sigma=abs(acos((l4.^2.+d.^2-l3.^2)/(2*l4*d))); %%angle between two bodies
keyboard
theta3(u,1)=tD + delta;

l1

[Morteza.Moslehi]

dear @Niels thank you so much to reply me soon but i didn't work..

[Niels]

if i am right u might alrdy know that l1 is a 1xnn vector ( where n is the number of rows from x)

try this

for nn=1:n
      theta3(nn)=tD(nn)+delta(nn)
end

by the way, i edited my answer after i saw that you also edited your question-> i saw that my first answer was incorrect

[Morteza.Moslehi]

that was right.. thank you so much for helping me.

just can you help me in following.

by doing this i have three answers for my theta 3 as see in following:

tD =

    3.1416
    3.1416
    3.1416
    3.1416
    3.1416
    3.1416
    3.1416

theta3 =

    4.7124

theta3 =

    4.7124    4.7124

theta3 =

    4.7124    4.7124    4.7124

theta2 =

     0

and i encounter with this Error for px:

??? Subscripted assignment dimension mismatch.

Error in ==> costtFunction at 46
    px(u,1)=real(P); %%real measure

would you please help me to cross this error too?

thank you for your favor

[Walter Roberson]

You have

l3=x(:,3);

so unless x happens to be a row vector, l3 is going to end up being a vector.

You have

alpha=acos((l3.^2+l5.^2-l6.^2)/(2*l3*l5));  % tha angle of coupler triangle

With l3 being a vector, alpha is going to come out as a vector.

You have

P=A*(l2*exp(1i*t2)+l5*exp(1i*(alpha+theta3(u,1)))); %%length of distance between input and specific point

with alpha being a vector, P is going to be a vector.

You then try to store real(P) into a single location.

[Morteza.Moslehi]

dear @Walter Roberson thank you for helping me just by doing this i encounter with this Error: ??? P= =

     0     0     0     0     
     0     0     0     0     
     0     0     0     0
     0     0     0     0

??? Subscripted assignment dimension mismatch.

Error in ==> costtFunction at 50 px(u,1)=real(P); %%real measure Inner matrix dimensions must agree. what should i do now?

[Niels]

i dont know how it was possible to get three results for theta3

either u get only 1, or 7, cause td has 7 rows and u didnt supress output of theta in the loop, but ok...

as Walter said, the problem is still the same as before: your result is a vector that u are trying to store in a single location:

u can either use the first value of P

px(u,1)=real(P(u);

or put P (and px, py) into the same loop as theta3, since it produces the same amount of results.

px(u,nn)=real(P(nn));

and if u didnt notice: u changed the way of calculatin theta3 in row 41 but in line 47

exp(1i*(theta3(u,1)))

you still use your old method, that mixed method is confusing.

Your code is a little bit messed up, since i guess the sizes of your values are not as they are expected to be. even if it might tun in the end it might not produce the result u hoped for.

[Walter Roberson]

Your u is a scalar, so px(u,1) refers to a single numeric location. real(P) has to give a single numeric result to be able to store it at px(u,1). I showed that your alpha is a vector so your P has to be a vector (or matrix) because you use all of alpha to create P.

You have to go back to your statement that creates P, and you have to look at size() of all of the variables that you use to create it.

[Morteza.Moslehi]

@Niels yeah you're right.i think my code is wrong but I'm newcommer in Matlab but i love learning that. so what should i do?? would you please help me?

although I followed your direction, i encountered again to this error:

??? Attempted to access tD(2); index out of bounds because numel(tD)=1.

Error in ==> costtFunction at 41 theta3(nn)=tD(nn)+delta(nn)

would you please help me?

[Niels]

thats weird, cause since l1 is a vector, r1, D and t_d(=angle(D)) and tD should be vectos as well. the error you get says that td contains only 1 value (has size 1x1)

can you uptdate to your current code? Maybe the full code, (and show me how what u set as n, where comes it from? is it the number of colums of x?) you cant just set n to what u like...

and let angle(D) be shown, do you know what angle does to a single value? i think angle does not do what you want it to as well. if you want the angle between r1 and r2, angle(r1-r2) would be wrong.

to make your script run and do what you want it to do would be far to difficult for me. there are 2 many things in your script that i do not know what they are for or what your goal is. in your if clause, the condition seems to compare 2 vectors. check with simple examples when the condition is true or false. 'If' works different for vectors than it does for scalars.

[Walter Roberson]

Strictly speaking, if works the same way for vectors and scalars: it is just that the way it works tends to surprise people when it is applied to vectors or matrices.

[Morteza.Moslehi]

@Niels sorry for being late.I've just arrived home from Hospital. the current code is about path generation of mechanism which designed just with 4 linkages.

l2 is input and can rotate for 360 degree and in following l3 and l4 rotate too. theta 2 in the angle between l2 and l1 that l1 is gound as you can see in code.

function cost = costtFunction(x)

global LowerBand UpperBand

l1=x(:,1); l2=2; l3=x(:,3); l4=x(:,4); l5=5 l6=4 theta1=0*pi/180; alpha=acos((l3.^2+l5.^2-l6.^2)/(2.*l3.*l5))

%% Gerashof principle %%

if l2+l3<=l1+l4; A=1; else A=0; end

R1=l1*exp(1i*theta1); %%length of 1 body

n=3; %% points u=0; %%initial point for creating loop

for t2=0:2*pi/n:2*pi %% theta2 and step which it goes u=u+1;

    R2=l2*exp(1i*t2); %%length of 2 body
    D=R1-R2; %%length of interact of 1 and 2 bodies
    d=norm(D); %%mesure of interact body
    tD=angle(D); %%angle of interact body wich called "D"

 delta=abs(acos((l3.^2.+d.^2-l4.^2)/(2*l3*d))); %%angle between two bodies
sigma=abs(acos((l4.^2.+d.^2-l3.^2)/(2*l4*d))); %%angle between two bodies
for nn=1:n
      theta3(nn)=tD(nn)+delta(nn)
end
    theta2(u,1)=t2 %%theta 2
%     theta3(u,1)=delta+tD %%theta3

% % P=A*(l2*exp(1i*t2)+l3*exp(1i*(theta3(u,1)))); %%length of distance between input and specific point P=A*(l2*exp(1i*t2)+l5*exp(1i*(alpha+theta3(nn))))

 px(u,nn)=real(P(nn)); %%real measure
   px(u,nn)=image(P(nn));%%imagine measure

end % G=sum(px)/n; %Sum of the x values and divide on total number of x % D=sum(py)/n; %Sum of the y values and divide on total number of y

% hold on % plot(px,py), xlabel('P_x (m)'), ylabel('P_y (m)'), title('Path of the point P') LowerBand = [4 2 1 3]; UpperBand = [5 2 3 9]; for ii = 1:length(LowerBand) x(:,ii) = LowerBand(ii)+(UpperBand(ii)-LowerBand(ii))*x(:,ii); end cost = ((x(:,1)) - (2)).^2 + ((x(:,3)) + (x(:,4))).^2;

i want to use this function to find best parameters of l1,l3 and l4. would you please guide me.?

Answer 3

[Morteza.Moslehi]

Dears @Niels and @Walter Roberson thank you so much for your favor. finally i found my mistake.. is should write my inputs like below to find scalar amount NOt vector.

cost=zeros(size(x,1),1); for ii= 1:size(x) l1=x(ii,1); l2=2; l3=x(ii,3); l4=x(ii,4); l5=3; l6=4;