dcegm.interfaces.interface

Interface functions.

Functions

get_n_state_choice_period(model_structure)

Get the number of state-choice periods from the model.

policy_and_value_for_states_and_choices(states, ...)

Get policy and value for a given state and choice vector.

value_for_state_and_choice(states, choices, params, ...)

Get the value function for a given state and choice vector.

policy_for_state_choice_vec(states, choices, ...)

Get the policy function for a given state and choice vector.

validate_stochastic_transition(params, model_config, ...)

Validate the exogenous processes in the model.

stochastic_transition_vec(state_choice_vec_dict, func, ...)

Evaluate the exogenous function for a given state-choice vector and params.

choice_values_for_states(value_solved, ...)

choice_policies_for_states(policy_solved, ...)

Module Contents

dcegm.interfaces.interface.get_n_state_choice_period(model_structure)

Get the number of state-choice periods from the model.

Parameters:

model (dict) – A dictionary containing model information. Must include ‘model_structure’ with a ‘state_choice_space’ key.

Returns:

A pandas Series with value counts of the first column of ‘state_choice_space’, sorted by index.

Return type:

pd.Series

dcegm.interfaces.interface.policy_and_value_for_states_and_choices(states, choices, params, endog_grid_solved, value_solved, policy_solved, model_config, model_structure, model_funcs)

Get policy and value for a given state and choice vector.

Parameters:

  • endog_grid_solved (jnp.ndarray) – Endogenous wealth grid for all states and choices.

  • value_solved (jnp.ndarray) – Value array for all states and choices.

  • policy_solved (jnp.ndarray) – Policy array for all states and choices.

  • params (dict) – Dictionary containing model parameters.

  • model (dict) – Dictionary containing model information and settings.

  • state_choice_vec (dict) – Dictionary containing a single state and choice.

  • wealth (float) – The wealth level at which to interpolate.

  • compute_utility (Callable) – Function to compute utility given the state, choice, and parameters.

  • second_continuous (float, optional) – An additional continuous state dimension. If provided, interpolation is done in two dimensions.

Returns:

A tuple of (policy, value) at the given state and choice.

Return type:

Tuple[float, float]

dcegm.interfaces.interface.value_for_state_and_choice(states, choices, params, endog_grid_solved, value_solved, model_config, model_structure, model_funcs)

Get the value function for a given state and choice vector.

Parameters:

  • endog_grid_solved (jnp.ndarray) – Endogenous wealth grid for all states and choices.

  • value_solved (jnp.ndarray) – Value array for all states and choices.

  • params (dict) – Dictionary containing model parameters.

  • model (dict) – Dictionary containing model information and settings.

  • state_choice_vec (dict) – Dictionary containing a single state and choice.

  • wealth (float) – The wealth level at which to interpolate.

  • second_continuous (float, optional) – An additional continuous state dimension. If provided, interpolation is done in two dimensions.

Returns:

The value at the given state and choice.

Return type:

float

dcegm.interfaces.interface.policy_for_state_choice_vec(states, choices, endog_grid_solved, policy_solved, model_structure, model_config)

Get the policy function for a given state and choice vector.

Parameters:

  • endog_grid_solved (jnp.ndarray) – Endogenous wealth grid for all states and choices.

  • policy_solved (jnp.ndarray) – Policy array for all states and choices.

  • model (dict) – Dictionary containing model information and settings.

  • state_choice_vec (dict) – Dictionary containing a single state and choice.

  • wealth (float) – The wealth level at which to interpolate.

  • second_continuous (float, optional) – An additional continuous state dimension. If provided, interpolation is done in two dimensions.

Returns:

The policy at the given state and choice.

Return type:

float

dcegm.interfaces.interface.validate_stochastic_transition(params, model_config, model_funcs, model_structure)

Validate the exogenous processes in the model.

This function checks that transition probabilities for each exogenous process are non-negative, sum to 1, and have the correct dimensionality. It raises a ValueError if any checks fail.

Parameters:

  • model (dict) – A dictionary representing the model. Must contain ‘model_funcs’ with ‘processed_stochastic_funcs’, ‘model_structure’ with ‘state_choice_space_dict’, and relevant ‘options’ keys.

  • params (dict) – Dictionary containing the model parameters.

Returns:

True if all exogenous processes are valid; otherwise, a ValueError is raised.

Return type:

bool

dcegm.interfaces.interface.stochastic_transition_vec(state_choice_vec_dict, func, params)

Evaluate the exogenous function for a given state-choice vector and params.

Parameters:

  • state_choice_vec_dict (dict) – Dictionary containing state-choice values.

  • func (Callable) – Stochastic state transition function to be evaluated.

  • params (dict) – Dictionary of model parameters.

Returns:

The exogenous process outcomes for the given state-choice combination and parameters.

Return type:

jnp.ndarray or float

dcegm.interfaces.interface.choice_values_for_states(value_solved, endog_grid_solved, state_choice_indexes, params, states, model_config, model_funcs)
dcegm.interfaces.interface.choice_policies_for_states(policy_solved, endog_grid_solved, state_choice_indexes, states, model_config)