2359 Views 11 Replies Latest reply: Feb 23, 2012 10:27 AM by ptc-3527070
10 posts since
Feb 14, 2012
Currently Being Moderated

## Urgent Optimization problem

I appreciate it if you could let me know if it is possible to minimize this in MathCAD. The problem is minimization of an error function with 24 constants(should be found by optimization) and 1 variable.

Error(a11,b11,d11,a12,b12,d12,a13,b13,d13,a22,b22,d22,a23,b23,d23,a33,b33,d33,a66,b66,d66,a77,b77,d77,eps)

So I want to find constant aij,bij and dij so that Error(eps) is minimum when 0<eps<0.1(at any eps), Is it possible to optimize this using Mathcad? If not would you please help me which math tool could possibly minimize this?

Thank you

Dariush

• 9,761 posts since
May 11, 2010
Currently Being Moderated
Feb 14, 2012 12:49 PM (in response to ptc-3527070)
Re: Urgent Optimization problem

It's not possible to say whether or not this is possible without more information. How is the function "Error" defined? Error between what and what?

• 9,761 posts since
May 11, 2010
Currently Being Moderated
Feb 23, 2012 1:26 PM (in response to ptc-3527070)
Re: Urgent Optimization problem

At least two of your parameters are not independent, so you have an infinte number of possible solutions. I believe that minimize has stopped at one of those solutions, which is very close to where you started. If you change the starting guesses you will find other, equally valid solutions.

Message was edited by: DanMarotta

• 9,761 posts since
May 11, 2010
Currently Being Moderated
Feb 23, 2012 1:26 PM (in response to ptc-3527070)
Re: Urgent Optimization problem

Yes that is true, But you see the problem I have right now is that I am fitting parameter "eps" along with other constants, although I want to keep it as a variable. What I am really looking for is a function Error(eps) that is always minimum within the range (0<eps<0.1) . Is that possible using mathcad?

Do you mean like in the attached worksheet?

Message was edited by: DanMarotta

• 35 posts since
Feb 10, 2012
Currently Being Moderated
Feb 15, 2012 11:43 PM (in response to ptc-3527070)
Re: Urgent Optimization problem

i think i  know what you want to do，but your worksheet are too complex，you may want to find the parameters in your function through minimizing the error，the error functio can be the error(between your function and the experimental data ) sum of squares.mathcad can do it.

my work may have something in common with yours.i give you two two suggestions,hope that it can do some help for you:

1.you have two many parameters,and your functions are very complex,you can use matrix instead of so many parameters;

2.the error function can be expressed like this:

you said that "What I am really looking for is a function Error(eps) that is always minimum within the range (0<eps<0.1)",i think it depends on whether the function and the data can be fitted well or not，if they are fitted very well，the Error maybe very small.

my english is poor,maybe i misurderstand you,wo can continue to discuss.

#### More Like This

• Retrieving data ...

#### Legend

• Correct Answers - 3 points