Minimize function with respect to multiple variables such that each variable has many variables?

Question

ABDULAZIZ ALTUN 2020-3-11

回答： ABDULAZIZ ALTUN 2020-3-11

I want to minimize a function with many variables

Suppose I have this function

function b = three_var(v,c)
	p=length(v);
	b = v(1)^2 + 2.5*sin(v(2)) -prod(v.*v)*c;
end

Then it is easy to do this

c=2;
fun=@(v)three_var(v,c);
v = [-0.6,-1.2,0.135, 1];
b_min = fminsearch(fun, v)

But suppose that the function is given by

function b = three_var(v,k,c)
	p=length(v);
	b = v(1)^2 + 2.5*sin(v(2)) +sum(k)-prod(v.*v)*c;
end

Where k is a vector. How can I find a way that minimizes my scalar b.

c=2;
fun=@(v,k)three_var(v,k,c);
v = [-0.6,-1.2,0.135, 1];
k=[3,2,1];
b_min = fminsearch(fun, v,k)

I am getting an obvious error because I am not using fminsearch in the right way but you know what i mean:

Error using fminsearch

Try Using Cell input

This did not work either, I tried rewriting the function as

function b = three_var(H,c)
	v=H{1};
	k=H{2};
	p=length(v);
	b = v(1)^2 + 2.5*sin(v(2)) +sum(k)-prod(v.*v)*c;
end

Then I wrote this

c=2;
H=cell(2,1);
H{1}=[-0.6,-1.2,0.135, 1];
H{2}=[3,2,1];
fun=@(v)three_var(v,c);
b_min = fminsearch(fun, H)

This gave me an error

Error using fminsearch (line 96)
FMINSEARCH accepts inputs only of data type double.

Can someone help me in this regards?

Answer 1

ABDULAZIZ ALTUN 2020-3-11

The best way around it

I could not find a nice solution to this:

I wrote the following function

function b = three_var(Input,c,nv,nk)
	v=Input(1:nv);
	k=Input(nv+1:nv+nk);
	p=length(v);
	b = v(1)^2 + 2.5*sin(v(2)) +sum(k)-prod(v.*v)*c;
end

Where nv and nk are the length of each of the input vectors

c=2
v=[-0.6,-1.2,0.135, 1];
k=[3,2,1];
Input=[v,k];
nv=length(v);
nk=length(k);
fun=@(v)three_var(v,c,nv,nk);
b_min = fminsearch(fun, Input)

this have solved my problem but I am sure there is a better way of doing this.