fminsearch with if statement

Hello all,

I need to optimise two parameters by minimising by objective function.

I've got an equation that has got two parts.

The parameters that I want to optimise are p0 and b0.

I already got the vectors obs and p.

The equation is:

If p < p0, then sim = p +b0*p0

if p>= p0, then sim = 5

Then I need to put my simulated values into my objective function, that looks like:

OF = (sim - obs)^2

So, now I want to minimise OF using fminsearch.

How can I do that, if the parameter that I am looking for is inside the constraint?

Thanks

Answer 1

[Matt J]

[xmin, fmin] = fminsearch(@(x) OF(x,obs,p), xguess);

function val=OF(x,obs,p)

    p0=x(1);
    b0=x(2);

    sim= (p +b0*p0).*(p<p0)+5.*(p>p0);

    val=(sim - obs).^2;

[Christina]

Hey Matt!

Thanks a lot for you reply, yes, that's mainly what I'd want to do.

However, the function option does not work for me, could you suggest how this could be written, without a function, i.e. as a simple script?

Thanks a lot.

[Christina]

Actually, I just found it.

Thanks a lot for your help!

Answer 2

[Matt J]

Looks like you're doing a first order spline fit with a single free knot, and therefore this FEX file might already do everything you need,

http://www.mathworks.com/matlabcentral/fileexchange/25872-free-knot-spline-approximation