Write a general purpose RK routine; here described in a matlab
context (but feel free to re-interpret into and program in any language environment)
[t_out,y_out,e_out] = rk ( ode_RHS, y_0, t_range, c, A, b1, b2 )
ode_RHS= function handle to right side f of the ODE y'=f(t,y)
y_0= initial value for ODE
t_range=the vector T_start:h:T_stop where h is the step size
c= the vector c in the butcher array that specifies the RK method
A= the matrix A in the butcher array that specifies the RK method
b1= vector in the butcher array that specifies primary RK method
b2= vector in the butcher array that specifies secondary RK method
t_out= the times where the solution was computed
y_out=the solution computed at the times t_out
e_out= the step-error estimated at the times t_out. Computed only when b2 is supplied.
Where is the step size h? You probably want to use something like
h_current_step=t_range(k) - t_range(k-1)
so that your code does not break if/when user passes in a non-uniform
e.g.t_range = [0 0.1 0.2 0.4 0.46 0.47 0.5 0.7 0.8 0.9 0.99 0.999 0.9999 1].
Two scripts are uploaded:
GenRK.m is the general Runge-Kutta script ...
The solution writes the general purpose of Runge-Kutta method. A description is provided in matlab.