pyperfectforesight.solve_steady_state#

pyperfectforesight.solve_steady_state(compiled_ss, params_dict, initial_guess=None, exog_ss=None)[source]#

Solve for the steady state using pre-compiled functions.

Uses the bundle produced by compile_steady_state_funcs() so that no symbolic work is repeated. Intended for parameter sweeps where the model structure is fixed but parameter values change across calls, and for computing terminal steady states at non-zero exogenous levels.

Parameters:

  • compiled_ss (dict) – Bundle returned by compile_steady_state_funcs().

  • params_dict (dict) – Parameter values. Must contain every parameter symbol detected during compilation (i.e., every key in compiled_ss['param_syms']).

  • initial_guess (ndarray, optional) – Initial guess for steady-state values (default: ones vector of length n_endo). When the model has multiple steady states or the solver struggles to converge, providing a better initial guess (e.g. a nearby known steady state) can help.

  • exog_ss (array-like of shape (n_exo,) or dict {str: float}, optional) –

    Steady-state values for the exogenous variables. Determines the long-run equilibrium at which the steady state is computed.

    • array-like — values in the same order as vars_exo passed to compile_steady_state_funcs().

    • dict — maps exogenous variable names (strings) to their steady-state values; missing variables default to zero.

    • None (default) — all exogenous variables are set to zero, which is the standard assumption for deviation models.

    Pass the last row of an exog_path here to compute the terminal steady state implied by the long-run exogenous level:

    >>> ss_terminal = solve_steady_state(compiled_ss, params, exog_ss=exog_path[-1])

Returns:

Steady-state values for each variable in vars_dyn order, wrapped in a SteadyState object that records the parameter values and exogenous levels used. Transparently usable as a plain ndarray via __array__.

Return type:

SteadyState

Raises:

ValueError – If params_dict is missing a required parameter symbol, or if exog_ss cannot be converted/flattened to a 1D array of length n_exo.

Examples

Standard use — exogenous variables at zero (deviation model):

>>> ss = solve_steady_state(compiled_ss, params)

Permanent technology shock — compute new terminal steady state at z=1.05:

>>> compiled_ss = compile_steady_state_funcs(equations, vars_dyn, vars_exo=['z'])
>>> ss_terminal = solve_steady_state(compiled_ss, params, exog_ss=np.array([1.05]))
>>> result = solve_perfect_foresight(
...     T, params, ss_terminal, model_funcs, vars_dyn,
...     exog_path=exog_path, ss_initial=ss_initial,
...     endval=ss_terminal,
... )
On this page

This Page