“…Its basis resides in the linear approximation φ(x 0 , p + h, t) ≈ φ(x 0 , p, t) + J h , where J = ∂φ/∂p is the Jacobian matrix at p, and h a small perturbation. Starting at an initial guess, the method produces a series of vectors vectors p (1) , p (2) , . .…”