Computation of convergence times

A custom program was written to do the following, starting from a regular 200*200 grid of initial conditions (for 2D systems), or a 50*50*50 grid (for 3D systems), with $\forall i\ne j, x_i\ne x_j$, to avoid reaching unstable steady-states: (1) integrate the system until a steady-state is reached (as defined by the sum of the absolute values of the derivative vector elements begin lower than $10^{-4}$) (2) start the integration again, with the same initial conditions, stopping when the system gets close enough to the previous steady-state (each variable with 10% of its steady-state value if it's not 0, lower than 0.15 if it is 0; moderate changes in these arbitrary values do not significantly affect the results). The stepsize of the Runge-Kutta algorithm was kept lower than 0.3.