Solving for the Global Nonlinear Saddlepath: Reverse Shooting vs. Approximation Methods

We present the blueprints for a set of innovative reverse shooting algorithms that trap the global saddle path in systems with 2-4 state variables. The solution procedure is built around a new distance mapping and refined simplex algorithms. Since the algorithms are completely reliable and always work in the same way, we have been able to develop canned problems that solve for the global nonlinear saddle path in any model with 2-4 state variables. The programs are written in the spirit of plug and play: the user types in the equations of the model and then waits for the solution.