How to code a math formula and passing it to a function

[Shadowrun]

hi!

I want to use a C++ function which receives a math function and executes minimization on it ( http://www.codecogs.com/d-ox/maths/o...ion/nelder.php) but I dont understand how to pass my formula as the first parameter.

There is even an example but I dont understand it...

double f(double *x) {
    double r = sqrt(x[0] * x[0] + x[1] * x[1]);
    return ABS(r) < 1E-12 ? 1 : sin(r) / r;
}

My formula is: http://personal.telefonica.terra.es/...dio1/text3.gif

This is what I've made:

for (i=1;i<=numberOfAnchors;i++)
{
   errorsquare+= pow( abs( x[i] - xestim + y[i] - yestim ) - distance[i]  ) , 2 );
}
errorsquare=errorsquare/numberOfAnchors;

Thanks.

[SlickHawk]

I'm not quite sure what you're asking. If it's about C++ syntax for referencing functions, this isn't really the forum for that, but whatever.

So, from what I've gathered, you need to make your squared error calculation a function and pass it into Maths::Optimization::nelder().

nelder() takes, as its first parameter, a function taking a pointer to a double (an array of doubles) and returning a double.

So, just transform for calculation into a function (let's call it "Foo"), where your unknowns are the parameters (it looks like x fills that roll, while y is known, based on your example), and pass it in by name as the first parameter of nelder().

double Foo(double * x)
{
     /*return your squared error based on x*/;
}

int main()
{
     Maths::Optimization::nelder(Foo, /*other params*/);
     return 0;
}

[Shadowrun]

You say that there is only 1 unknown (x), but there are 2 ('xestim' and 'yestim'). Research paper shows "P" (point) as the unknown, so I guess it should be (x,y) 2 unknowns....

The C++ function that I'm trying to use says: "This method performs the minimization of a function with several variables..."

So my question is precisely how I have to code my formula so I can pass it to the C++ function, i.e.

/*return your squared error based on 'xestim' and 'yestim' */