2
$\begingroup$

I need to solve a differential-functional equation:

$\partial_t x_m(t,s) = \sum_n A_{mn} x_n(t,s) + \int_0^t \sum_n \sum_{n'} B_{mnn'}(t,s') x_n(t,s) x_{n'}(t,s') ds'$

with $t > s$ and initial condition $x_m(t,t) = C_m$.

What is the best numerical procedure?

$\endgroup$

2 Answers 2

0
$\begingroup$

Huh, that's one strange equation. Where does it come from? Are you sure there's not a simple form? (in particular, one you can differentiate to give a simple differential equation) Anyway, look up litterature for integro-differential equation. Or just bruteforce it with an Euler method and piecewise constant integration, if you don't need high precision.

$\endgroup$
3
  • $\begingroup$ Pls note that the differentiation is in the first parameter, and integration over the second. $\endgroup$
    – Katastrofa
    Dec 4, 2011 at 11:23
  • $\begingroup$ Right, sorry. So your equation just doesn't make any sense, because there is no constraint on the s-evolution of x. Think first of the simplest case : in one variable, A = 1, B = 0, $\partial_t x(t,s) = x(t,s)$. Impose an equation for $\partial_s x$ or a constraint. $\endgroup$ Dec 4, 2011 at 12:00
  • $\begingroup$ There is an initial condition for x(t,t). $\endgroup$
    – Katastrofa
    Dec 4, 2011 at 17:13
0
$\begingroup$

I have solved it in a trivial way: using Euler discretization in the $t$ direction and trapezoid quadrature in $s$ direction, with equal steps. Works quite well for the data I use.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.