Derive function handle with a vector input argument

when I run this piece of code

q=[3 2 2 3];
f=@(p,x,y) p(1)*x.^2+p(2).*y.^2+p(3)*x.*y+p(4);
syms 'x'
df=diff(f,x);
dfdx=matlabFunction(df);
 A=feval(f,q,2,2);

I get an error says

 Index exceeds the number of array elements (1).

What I am trying to do is to cluster my equation coefficients into one vector to make my code compact. In my real application I have 15 coefficients. Any advice if this process is possible?

Answer 1

[Walter Roberson]

q=[3 2 2 3];
f=@(p,x,y) p(1)*x.^2+p(2).*y.^2+p(3)*x.*y+p(4);
syms p [1 4]
syms x y
df=diff(f(p,x,y),x)

df = 

dfdx = matlabFunction(df, 'vars', {p, x, y})

dfdx = function_handle with value:
    @(in1,x,y)in1(:,1).*x.*2.0+in1(:,3).*y

A = feval(f,q,2,2)

A = 31

[Majeed]

Thank you very much!

Answer 2

[Paul]

Hi Majeed,

Maybe you're looking for something like this?

q = [3 2 2 3];
p = sym('p',[1 4]);
syms x y
f = p(1)*x.^2+p(2).*y.^2+p(3)*x.*y+p(4)

f = 

df=diff(f,x)

df = 

A = subs(f,[p x y],[q,2,2])

A = 

31

Answer 3

[Matt J]

If you have the Deep Learning Toolbox, you can do automatic differentiation of vector-valued dlarrays,

p=1:15;
x0=dlarray(rand(size(p)));

[fx,gradfx] = dlfeval(@(x) somefunc(x,p),x0)

fx = 
  1×1 dlarray

   68.8215

gradfx = 
  1×15 dlarray

     1     2     3     4     5     6     7     8     9    10    11    12    13    14    15

function [y,dydx]=somefunc(x,p)
 y=p*x';
 dydx=dlgradient(y,x);
end