Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > comp.soft-sys.math.mathematica > #4825

Re: two coupled differential equations

From "Kevin J. McCann" <Kevin.McCann@umbc.edu>
Newsgroups comp.soft-sys.math.mathematica
Subject Re: two coupled differential equations
Date 2011-09-08 09:26 +0000
Organization Steven M. Christensen and Associates, Inc and MathTensor, Inc.
Message-ID <j4a1oa$40e$1@smc.vnet.net> (permalink)
References <j3t57m$56h$1@smc.vnet.net>

Show all headers | View raw


First, the de for x1 is not coupled to that for x2; so, solve it 
separately. You can do this and then find the optimum k1 and initial 
condition on x1 (you didn't specify this). Then proceed to that for x2 
(numerical solution only). This will not necessarily give you the global 
best fit to the parameters, but at least it is a start.

I will send you a notebook separately that gives my attempt.

Kevin

On 9/3/2011 8:06 AM, Mark.Musters@leadpharma.com wrote:
> Hi
>
> I have two coupled differential equations, for which I would like to
> derive the parameter values by fitting them on a data set.
>
> The differential equations have the following form:
>
> x1'[t] == -k1*x1[t]
> x2'[t] == k2*x1[t]-k3*x2[t]/(k4+x2[t])
>
> The experimental time-series data has the following form
> t== {0,1,4,8,12,24}
> x1 == {10,9,5,3,1,.1}
> x2 == {2,4,5,7,9.1,11}
>
> The examples I can find with FindFit are aimed at a single differential
> equation, so I was wondering how I can assign the data to the
> corresponding state variable and perform a parameter estimation to get
> an estimate for parameters k1, k2, k3 and k4.
>
> It would help me a lot.
>
> Many thanks in advance,
> Mark

Back to comp.soft-sys.math.mathematica | Previous | Next — Previous in thread | Find similar | Unroll thread


Thread

two coupled differential equations Mark.Musters@leadpharma.com - 2011-09-03 12:06 +0000
  Re: two coupled differential equations "Kevin J. McCann" <Kevin.McCann@umbc.edu> - 2011-09-08 09:26 +0000

csiph-web