how to input a function as a parameter of a function

Hi there,

Here I have the following recursive code fMin1D to find the local minima of a user-defined function f=fun(x).

fun(x) is able to evaluate a sample array x =[x1 … xn] and return values f =[f1 … fn] .

function x0=fMin1D(x1,xN,N,eps,x0,fun)
% samples to be scanned
Neps=1+ceil((log(xN)-log(x1))/log(1+eps));
n=max(4,min(N,Neps));
x=x1*(xN/x1).ˆ([0:n-1]/(n-1)); f=feval(fun,x);
% minimum search via direct comparison
L=f(1:n-1)<=f(2:n); I=find(L(2:n-1) & ~L(1:n-2))+1;
% recursive search or update of already found solution x0
if n<Neps, for j=1:length(I),
        x0=fMin1D(x(I(j)-1),x(I(j)+1),N,eps,x0,fun); end
else x0=[x0,x(I)]; end

Since fun is also a function, how can I input it into the brackets of fMin1D as a parameter needed by fMin1D?

Many thanks!

Cheers

Answer 1

[Sam Chak]

Hi @Tony Cheng

Something like this:

%% Script to call 'fMin1D'
x0  = [1, 1];
[x, fval] = fMin1D(x0, @quadfcn)

Local minimum found.

Optimization completed because the size of the gradient is less than
the value of the optimality tolerance.

x = 1×2

0 0

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

fval = 0

%% Minimizer
function [x, fval] = fMin1D(x0, fun)
    [x, fval] = fminunc(fun, x0);
end

%% Quadric surface
function y  = quadfcn(x)
    y = x(1).^2 + x(2).^2;
end

[Tony Cheng]

Hi Sam,

I guess what u mean is, in the main .m file, when I call the fMin1D function, @ is added before the user-defined function fun. But in the file where the function fMin1D is defined, @ is not needed before fun.

Is my interpolation correct?

Cheers

Tony

[Sam Chak]

Yes, @Tony Cheng, your understanding is correct.

[Tony Cheng]

Hi Sam, thx so much for your sincere help! I make it now!

Answer 2

[Walter Roberson]

Use @fun as the parameter. For example

X0 = fMin1D(.5, 17, 32, 1e-10, 0.8, @fun)

[Tony Cheng]

Hi Walter, pls see the attached screenshot:

there is a red waved line under the @ symbol...