| T O P I C R E V I E W |
| kdausman |
Posted - 09/14/2007 : 5:49:36 PM
Origin Version (Select Help-->About Origin): OriginPro 7.5 SR6
Operating System: Windows XP
Short version of question: How do I generate a fitting function that is the solution to a small-to-medium-size initial value simulation?
Long version of the question: Let's say that I have Concentration vs. Time data that I want to model by a set of kinetic rate equations that I have in differential form (i.e., a set of differential equations). Outside of Origin I would do this by numerically integrating the set of rate equations using the 4th-order Runge-Kutta method (an Excel add-in to do this is found here: http://www.chem.mtu.edu/~tbco/cm416/RK4_v3_0.html), and would then systematically vary the input parameters (rate constants and initial concentrations) to achieve the best fit. Is there a way to do this in Origin? |
| 6 L A T E S T R E P L I E S (Newest First) |
| kdausman |
Posted - 10/13/2007 : 6:27:05 PM
I agree, it would be an excellent addition. As it is, however, I have taught myself enough Mathematica that I think I can now do it there. I'd prefer to do it in Origin, certainly, because I really like the interactive fitting, but for the time being I am getting done what I need to. |
| matsati |
Posted - 10/13/2007 : 3:46:43 PM
This would be an awsome feature if it was included in a future release! |
| larry_lan |
Posted - 10/08/2007 : 06:10:58 AM
I think now Origin can't fit such kind of differential equation.
Thanks
Larry
OriginLab Technical Services |
| kdausman |
Posted - 10/05/2007 : 5:05:45 PM
Thank you for the help. Unfortunately, I don't see how to apply that to my situation, because my problem isn't a simple integral. Let me give you a specific example.
Suppose that I have the following two differential equations:
dA/dt = -k1*A - k2*A^2 - k3*A*B
dB/dt = k1*A + k2*A^2 + k3*A*B
The data I am trying to fit is A (dependent variable) vs. t (independent variable). B is a hidden dependent variable. My fitting parameters are k1, k2, k3, A0, and B0, where A0 and B0 are the values of A and B at t=0.
So, to restate my original question, what I want to do is start with values of A and B, do numerical integration for incremental values of t according to the equations above, and generate a simulated plot of A vs. t, which I then fit to experimental data by varying the ks and A0 and B0. I don't see how to do that from the link you provided. Can you help further?
Thanks!
Kevin |
| Deanna |
Posted - 09/23/2007 : 10:59:02 PM
Hi,
It is possible to perform fitting with a function that involves integration. You may see instructions in this page:
http://www.originlab.com/www/support/resultstech.aspx?ID=121&language=English&Version=7.5
Deanna
OriginLab Technical Services |
| gorkin |
Posted - 09/23/2007 : 09:37:47 AM
Yes there is. You can do numerical integration with origin, but implementation is much more difficult than with other software like Scientist. I would like to see this sort of functionality in origin in the future. |
