Calling vi from formula node

[altenbach]

Ken,

Have a look at my new example (and related discussion higher up in the same thread). You could probably modify it so it works with your model.

Message Edited by altenbach on 04-08-2005 06:02 PM

[altenbach]

---

@Ken Pietrzak wrote:
I've attached the error function that I'm trying to minimize.

---

Ken,
Do you have the formula for the functions g and h?
x'k = g(xk, yk, K ,xc, yc)
y'k = h(xk, yk, K ,xc, yc)

[Ken Pietrzak]

Thanks guys. I really appreciate the help. I've attached a more thurough definition of my error function as well as a VI that executes it. The params array input in the error function contains the initial search parameters. K1 and K2 should be quite small - 0.1 at the most and I expect them to be much smaller. Xc and Yc should vary by less than a few hundred for the initial params. The theta's and p's should not change much at all.

Ken

[Ken Pietrzak]

Here's the error function:

[Ken Pietrzak]

I took altenbach's advice and modified NI's Lef-Marq VI to solve for z = F(x,y) where z always equals 0. I removed the squaring and summations and instead let Lev-Marq use the chi-square approach to solve for the params. I find that it converges to expected values quite quickly. Thanks again for your help.

[altenbach]

Ken,

It is great to hear that it works!
Sorry I haven't had time to look into this myself, but it seems you're all set. 🙂

Happy wiring...