Abnormal termination of minimization : reasons?

Dear All.

I am fitting some function at data for range 2 to 200, but it failed. When i change the range to 2 to 108 it gives me nice fitting. But just after 108 the functions fails to fit with message “Abnormal termination of minimization” .
So i tried to understand the output and reason for this, what i found is

  1. till range 108 function is fitting nice with status:
FCN=62.2591 FROM MIGRAD    STATUS=CONVERGED     313 CALLS         314 TOTAL
                     EDM=8.22293e-09    STRATEGY= 1      ERROR MATRIX ACCURATE 
  EXT PARAMETER                                   STEP         FIRST   
  NO.   NAME      VALUE            ERROR          SIZE      DERIVATIVE 
   1  c            9.63910e+02   1.76001e+01   2.16986e-02  -1.14205e-05
   **2  q            1.35079e+00   3.94012e-02   9.05930e-02   5.74516e-07**
   3  k            1.34253e+00   3.95433e-02   2.60646e-05   6.40928e-03
   4  nV           5.51199e+00   1.23216e-01   5.47718e-05  -3.20463e-03
   5  nV0          1.15495e-01   2.67276e-01   3.67715e-03   4.71362e-05

So looking closely at this output i found the First derivative of 2nd parameter is very small e-07.

  1. when i fit the data for full range 2 to 120 it gives me output as
FCN=6.3077e+306 FROM MIGRAD    STATUS=FAILED        284 CALLS         285 TOTAL
                     EDM=inf    STRATEGY= 1  ERROR MATRIX UNCERTAINTY 100.0 per cent
  EXT PARAMETER                APPROXIMATE        STEP         FIRST   
  NO.   NAME      VALUE            ERROR          SIZE      DERIVATIVE 
   1  c            1.00000e+00   1.41421e+00  -0.00000e+00   0.00000e+00
   2  q            1.32690e+00   3.57038e-02  -0.00000e+00   0.00000e+00
   3  k            2.50000e+01  1.74468e-152  -0.00000e+00  1.52685e+305
   4  nV           6.00000e+00   1.41421e+00  -0.00000e+00   0.00000e+00
   5  nV0          2.80000e-01   2.46491e-01  -0.00000e+00   0.00000e+00

So it is clear that step size and its derivative is playing role in this failure and success of function.

Could some one please elaborate it more that how this works or any document/info available on this from which i can read and understand the reason better

Thanks and Regards
Sandeep

The first thing to try is to set “reasonable” initial values for all parameters of your fitting function. They should be set as close as possible to “real values”, otherwise the fit procedure may fail to find the minimum (and you will get different failure messages, including “abnormal termination”).

HI,

It looks to me the function (FCN) is having a non-sense value, ~ E306. So probably increasing the fitting range you might have some singularity. You should check this

Lorenzo

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.