Stochastic Differential Equation Simulation Using simByEuler

I have a system of stochastic differential equations that I want to solve numerically. This system consists of 6 equations, so the drift term is a 6*1 vector and the diffusion is 6*6 symmetric matrix. I can not solve it analytically and all I need is to solve the system numerically using simByEuler command (found in financial toolbox). The problem is that I got complex value solutions which I do not want because the model I am working on is dealing with orientation of particles in fluid. This orientation should be real values so that it is easy to implement and test experimentally. Is there any suggestion how to avoid complex solutions.

Answer 1

[Torsten]

My guess is that your equations contain terms with like expr^a, log(expr), sqrt(expr) or something similar where expr becomes negative during integration.

This will result in complex number solutions.

Best wishes

Torsten.