occbin fixes and two small new features

See merge request Dynare/dynare!1883

1) algo_truncation: sets the number of iterations for a truncated guess-verify algorithm (i.e. if max_iter<=algo_truncation, no error is triggered on output, but the user is happy with the last regime in the algorithm)
2) reset_regime_in_new_period: is set to true, it resets the guess regimes to unconstrained when a new shock arrives, instead of starting with a guess regime consistent with the one identified in previous time periods. this sometimes allows more robust convergence/identification of the regimes

When the VAR auxiliary model has a constant.

make header names compatible with earlier versions of MATLAB (e.g. it would crash with MATLAB 2017b)

1) if not using relaxation, initial guess regime in verify step(occbin simul) is set equal to the guess regime used to updated states and shocks
2) fixed treatment of periodic solution in algo=3, now harmonized to algo=1

Initialize state space matrices with occbin smoother, when realtime regimes are computed, otherwise they are set to 0 for t<first_period_occbin_update!

with occbin smoother do not overwrite updated variables that have been already computed with reduced state space

Ref. !1882

Ref. !1881

— Fix preprocessor crash when a [static] equation contains a “var_expectation” operator
– Occbin: handle external functions in “occbin_constraints”
— Occbin: add more sanity checks on expressions in “occbin_constraints” block

Ref. #569

Some fields have also changed under M_.occbin.

Ref. #569

license.txt: correct Occbin license information

See merge request Dynare/dynare!1880

[skip CI]

Marco Ratto authored May 17, 2021 and Sébastien Villemot committed Jul 16, 2021

Syntax is not yet finalized (see preprocessor#68).
Documentation still to be done.

Ref. #569

Ref. #1785

Just add option `structural` to the `var_model` command.

NLS estimation does not work if the equation has expected variables.

Previous definition could detect lags in an equation without any lag, for
instance if we have the following expression in the RHS: +(-1)*x.

Ref. #1798.

See example in tests/estimation/example_nls.mod.

initial_estimation_checks.m: error out if data is complex

See merge request Dynare/dynare!1879

Prevents cryptic message that likelihood became complex

Ref. #1788

Previously any nonzero mean exogenous variable had to be centered on
the growth rate (first difference) of the (PAC's LHS) endogenous
variable.

In PAC equations with optimizing and rule of thumb agents (division
by lambda was missing). Also add comments.

These command solve the problem where agents think they know perfectly the
future (they behave as in perfect foresight), but make expectation errors.
Hence they can potentially be surprised in every period, and their expectations
about the future (incl. the final steady state) may change.

Currently the sequence of information sets needs to be passed through a CSV
file. Another interface may be added in the future.

The algorithm uses a sequence of (true) perfect foresight simulations (not
necessarily as many as there are periods, because if the information set does
not change between two periods, there is no need to do a new computation).

There are two possibilities for guess values:
— the default is to use the initial steady state for the simulation using the
  first-period information set; then use previously simulated values as guess
  values
— alternatively, with the terminal_steady_state_as_guess_value option, use the
  terminal steady state as guess value for all future periods (this is actually
  what the “true” perfect foresight solver does by default)

In several places, there was only a general link to the old Dynare wiki. This
can be confusing, because we no longer use this wiki. Rather use specific links
to the pages that were not migrated to the new wiki.

[skip ci]

— New “structural” option to “var_model” (#1785)
— Remove support for obsolete syntax: var_model(order = 2) X Y Z;
— “var_model” statement: make “eqtags” option mandatory