pythonodescikits

How to feed Jacobian to scikits.odes.odeint

I am using scikits.odes.odeint to solve a relatively large system of ODEs (40 variables, 100 kinetic parameters). I want to speed up the process by supplying a jacobian to the solver. However, I do not know how to best define the jacobian for this purpose and supply it to the solver.

Solution

  • This is a very basic symbolic Jacobion but it works for most ODE solvers.

    import sympy as sp

def rhs(t, y):
return [y[0],2*y[1]]


def jacob(t, y):
  ydot = rhs(t,y)
  J =  sp.Matrix(ydot).jacobian(y)
  J_func = sp.lambdify((t, y) , J)
  return J_func

    There is a more in depth guide here.