Syms function optimization. How to solve function (NLE) for minimum value?

I have derived the equation following syms of variable D s n n_pass as the following function (fun).

fun=100000*(250000000/(549*n*pi*(2*s + 1/1000)) - 2000*n_pass + (625000000*s)/(157*n*(D^2 - 5312662293228351/147573952589676412928)))^2 + 100000*(48000000000/(7*n_pass) - 1000000000000*n*s*((157*D^2)/100 - 834087980036851107/14757395258967641292800))^2 + 7850*D^2*n*n_pass*(2*s + 1/1000) + (102400*n*(2*s + 1/1000))/(627*n_pass*pi)

I have to minize the following function. How to do it using MATLAB optimization functions. Please help?

[Torsten]

Minimize with respect to all 4 variables D, s, n and n_pass ?

Best wishes

Torsten.

[Akshayy Garg]

Yes. The optimum minimum solution for the above. I tried Lagrange its diverging. I tried DEA but I get multiple solutions every time. So I was looking for MATLAB optimization functions. But in that variables are written in x(1), x(2) format. So how can I convert D, s or any syms variable into x-matrix format. Or if any other technique would help?

Regards Akshayy Garg

[Walter Roberson]

matlabFunction with the vars option. List the variables within a vector within a cell array

   'vars', {[D, s]}

[Akshayy Garg]

I know that. But I alraedy have the equation with the syms variable in U as follows: U =

100000*(250000000/(549*n*pi*(2*s + 1/1000)) - 2000*n_pass + (625000000*s)/(157*n*(D^2 - 5312662293228351/147573952589676412928)))^2 + 100000*(48000000000/(7*n_pass) - 1000000000000*n*s*((157*D^2)/100 - 834087980036851107/14757395258967641292800))^2 + 7850*D^2*n*n_pass*(2*s + 1/1000) + (102400*n*(2*s + 1/1000))/(627*n_pass*pi) I want to replace D,s,n,n_pass in the equation to x(1),x(2),x(3), x(4) directly. How to do that?

[Walter Roberson]

subs(U, [D, s, n, n_pass], x(1:4))

Answer 1

[Torsten]

fun=@(D,s,n,n_pass) 1000*....;
x0=[1 1 1 1];
[x,fval,exitflag,output] = fminunc(@(x)fun(x(1),x(2),x(3),x(4)),x0)

Best wishes

Torsten.

[Akshayy Garg]

I got my earlier function stored in var U.I tried this way: U=1000*....; % I cannot write this directly in fun because I get the values from other equations also. This is the final objective function fun=@(D,s,n,n_pass)U x0=[10e-3 1e-3 5e3 20] [x,fval,exitflag,output]=fminunc(@(x)fun(x(1),x(2),x(3),x(4)),x0)

Did not work?

Regards Akshayy

[Torsten]

Why don't you use the code from above ?

fun=@(D,s,n,n_pas) U won't work, and you forgot the "x0" in your call to "fminunc".

Just do the following:

1) Remove all "syms" declarations in your code.

2) Copy the code from above and insert your fun-definition in the line fun=@(D,s,n,n_pass) 1000*...

3) Run the script.

Best wishes

Torsten.