We present an algorithm and software routines for computing nth-order approximate solutions to dynamic, discrete-time rational expectations models around a nonstochastic steady state. We apply these routines to investigate the optimal monetary policy with commitment (and from a ``timeless...