rk_adaptive employs the runge-kutta method for numberically solving differential
equations, but other than rk it tries to find a sensible step size, which reduces
both runtime and amount of data produced.
startval, endval, params1...)
Tries to solve the ODE whose derivates are defined by expr to the variables
vars, starting at their initial values vars_initval.
The independent variable (in physics: normally time) is var, being stepped from
startval to endval.
Sometimes it speeds up the calculation to use a floating-point number as
startval, as using floars will prevent all results from becoming endless rational
numbers.
The following optional parameters are accepted:
maxstep=num: The maximum step size to be used.
minstep=num: The minimum step size to be used.
timestep_initial=num: The initial guess for the optimum time step to start with.
maxabserr=num: The maximum absolute error of the resulting curves.
maxrelerr=num: The maximum relative error of the resulting curves as a
fallback for variables with big values for which maxabserr might be too sensitive.
See also rk.
Example:
(%i1) pnts:rk_adaptive(-1/10*x,x,1,t,0,100) $