cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
Showing results for 
Search instead for 
Did you mean: 

Community Tip - Learn all about PTC Community Badges. Engage with PTC and see how many you can earn! X

Minerr/nonlinear optimizing

ycho-3
1-Visitor

Minerr/nonlinear optimizing

I tried to minimize the objective function above, using Minerr function in Mathcad.

and i put the guess value of 18 variables

and then finally i could get the  optimized value of them

but i am not sure the work i done is right or wrong, so i need a check.

can anyone can check my work?

2 REPLIES 2
Werner_E
25-Diamond I
(To:ycho-3)

If you want to ask a question, its better to open a question rather than creating a document. You can attach files to a question as well.

Here we can only add comments but are unable to provide files.

At first it looks rather strange that you write 1=1, etc. This does not make any sense - delete those regions.

I haven't looked at the upper parts of your worksheet how your functions and vectors are defined and derived, but concerning Mathcad syntax you had done it right. You end up with a set of 18 values which makes your function "function" yield 3.189*10^4 as result. Not zero and I am not sure if this is OK for your application.

In Mathcad you can chose among several algorithms used by right clicking the word "minerr". The default is Levenberg-Marquardt and usually thats a good choice, but sometimes other algorithms may give better results. I tried with your values and here the simple Quasi-Newton seems to give the best result (1512).

You may also consider to use "Minimize" instead of "Minerr". Using the same guess values we get this:

Of course guess values can influence the outcome as well and you may experiment by using the outcome of  Minimize as guess for the MinErr algorithm. Not sure if this would give you better results.

Regards

Werner

RichardJ
19-Tanzanite
(To:ycho-3)

As already noted, a document is not a good way to post a question, because we can't attach files to our comments. See my answer here: Minerr/nonlinear optimizing

Announcements

Top Tags