Take the 2-minute tour ×
Mathematica Stack Exchange is a question and answer site for users of Mathematica. It's 100% free, no registration required.

Is it possible to derive a cross-correlation function between a stochastic variable and a state-variable in an SDE, such as for the simple model here, or better between two state variables in a two variable (plus one stochastic variable) model:

sys = ItoProcess[{\[DifferentialD]n[t] == (ibar + rP p[t] - lN n[t] - u1 n[t]) 
\[DifferentialD]t + sd \[DifferentialD]w[t], 
\[DifferentialD]p[t] == (u1 n[t] - (rP + lP) p[t]) \[DifferentialD]t}, 
{n[t], p[t]}, {{n, p}, {n0, p0}}, t, w \[Distributed] WienerProcess[0, sd]];

I am interested in the long-term limit of correlations between variables, starting from initial conditions that are equal to the means. For instance for the one variable case, I am able to compute expressions for the mean and variance (and autocorrelation, but not shown):

proc = ItoProcess[\[DifferentialD]n[t] == ( ibar - q n[t]) \[DifferentialD]t + 
       sd \[DifferentialD]w[t], n[t], {n, ibar/q}, t, 
       w \[Distributed] WienerProcess[0, sd]](*this is white noise*);

Mean[proc[t]]
Variance[proc[t]]

Out[2]= ibar/q

Out[3]= (E^(-2 q t) (-1 + E^(2 q t)) sd^4)/(2 q)

 Limit[E^(-2 q t) (-1 + E^(2 q t)) /. q -> 1, t -> Infinity]

Out[4]= 1

However, large outputs are generated for the two variable case. Is there a way to solve for variances (mean can be computed easily on paper) and the cross correlation between any of the variables (ibar + sd dw[t]/dt, n[t], p[t]) in the limit shown above?

Is a symbolic approach possible? If not, are there approaches to approximate a function, either from numerical simulations, Taylor expansion or another technique? Suggestions and links to any existing resources are appreciated.

share|improve this question
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.