![[FRB logo]](../bitmaps/sm_head2.gif)
|
The Anderson-Moore Algorithm AIM Home Page |
![[AIM logo]](../bitmaps/aimLogo1.0.gif)
Accelerating Non Linear Perfect Foresight Model Solution by Exploiting the Steady State Linearization
Linearizing non linear models about their steady state makes it possible to use the Anderson-Moore Algorithm(AIM) to investigate their saddle point properties and to efficiently compute their solutions. Using AIM to check the long run dynamics of non linear models avoids many of the burdensome computations associated with alternative methods for verifying the saddle point property. In addition, for models that have the saddle point property, AIM provides a set of terminal conditions for solving the non linear model that work better than the traditional approach of setting the end of the trajectory to the steady state values. Furthermore, the asymptotic linear constraints can also generate initial conditions for the solution path that are better than initializing the solution path to the steady state values. Using the improved asymptotic constraints typically halves the computational burden associated with solving the nonlinear problem.