all-at-once fit - problems with constraints


I would like to make multivariate all-at-once fit with many coefficients (K[0], K[1], ..., K[n]). I need to constrain coefficients using absolute value of the difference between subsequent coefs:

abs(K[i+1] - K[i])<0.1.

Is there a possibility to realize such kind of constraints? I found in the Igor manual that only linear combinations of coefs are allowed?

I would be grateful for an answer.


In reply to by thomas_braun


I use the following command to make a fit:

FuncFit/H=hold_str/NTHR=0 voigt_2D parametry  data_l /X={skala_x_l,skala_y_l}/C={w_m,l_w} /D=data_kopia

where w_m is matrix with constraints and l_w is a vector with limits.


(I'm sorry, but previously I deleted the content of the post trying to edit it.)

Your constraint is linear except for the call to abs(). You can achieve that constraint with two constraints:

K[i+1]-K[i] < 0.1
K[i+1]-K[i] > -0.1