Get Help

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

- Community
- :
- PTC Mathcad
- :
- PTC Mathcad
- :
- Maximum parameters possible to minimize?

Options

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

Highlighted
##

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Notify Moderator

10-24-2012
11:28 AM

10-24-2012
11:28 AM

Maximum parameters possible to minimize?

Hi,

I am using Minimize trying to find some 200 parameters that result in a minimum value for a function. Now I noticed that the last 80 parameters (after parameter # 120) do not not change during the optimization process. Now I wonder if 120 is the maximum parameters that Minimize function accepts? I am using version 14.0 M011.

Thank you in advance

Dariush

Labels:

4 REPLIES 4

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Notify Moderator

10-24-2012
12:27 PM

10-24-2012
12:27 PM

Re: Maximum parameters possible to minimize?

From the Mathcad 14 Help ...

- Solve Blocks can solve systems of up to 400 variables when using the Quasi-Newton or Conjugate-Gradient solvers. The Levenberg-Marquardt solver is unconstrained, and will solve any size system. Linear systems can have up to 8192 constraints, and nonlinear systems up to 200 constraints. When you solve for complex variables, Mathcad treats the real and imaginary parts as separate variables in the algorithm, increasing the overall variable and constraint count for the problem.

Stuart

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Notify Moderator

10-24-2012
01:36 PM

10-24-2012
01:36 PM

Re: Maximum parameters possible to minimize?

The function Minimize and others with variable number of arguments (find, min, max, mean etc) can have not more 50 arguments. We can use at more arguments a vector-argument.

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Notify Moderator

10-24-2012
01:54 PM

10-24-2012
01:54 PM

Re: Maximum parameters possible to minimize?

Stuart and Valery,

Thank you for your replies. To make thos more clear, I have a function which has 10 vectors as variables. Five of them have 35 Unknown and the other five have 5 unknow that has to be found through Minimize. There are 40 contraints. I took a print screen of how it looks like:

So here A,Ap,... are vectors with 35 unknowns and B,Bp,...are vecotrs with 5 unknowns. K,Kp,...are function of B vectors and there are 40 of them. I am using Nonlinear conjucage gradient algorithm. What I am seeing here is that the value of A,Ap,App,B,Bp,Bpp (120 uknowns in total) are being optimized but the value of Appp,Bppp,Apppp,Bpppp does not change at all after optimization is done (so the values remain the same it was before optimization).

Based on what you mentioned there should n't be any problems. Am I correct? What could be wrong here. Please note that if I change the initial guess for Appp,Bppp,Apppp,Bpppp the values of the function "Error" that I am trying to optimize changes.

Thank you

Dariush

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Notify Moderator

10-24-2012
03:03 PM

10-24-2012
03:03 PM

Re: Maximum parameters possible to minimize?

I think it will be better to use not Minimize but MinErr function.