Why are the results calculated from the function "fmincon" not stable?

Hallo everbody,

Recently I am doing the fit about burning velocity of fuels, which is one of the most important properties in the combustion field. The approximation formula is as follwed:

In this formular there are six coefficients, namely F,m,G,n,E,B. And the variables: p,Y,Tu,Tb are already given.

Then I use the mathematic function "fmincon" to calculate these six coefficients. I have set the initail coefficients value and the highst and lowest coefficients value. But the results calculated are not stable. The coefficients value are always changed.

I do not know where the problem is.

[Matt J]

Are m and n supposed to be integers? FMINCON will not restrict its search to integers. You would need ga() for that.

Answer 1

[Anthony]

to my understanding, all fmin*** functions are for local minimum only. If your function has multiple local minimizers, then the solution will depend on the initial condition.

[kang wang]

Hello Anthony! First, thank you for your answer. In my case, the local minimizer is the least-square method as followed:

Futhermore, I used the "globalresearch" and "multistart" to run this script. And the results are still not stable.