stepan
/
dynare
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Add headers to various functions

time-shift
Johannes Pfeifer 2016-07-05 19:41:24 +02:00 committed by Stéphane Adjemian (Lupi)
parent dfd44c58d6
commit 1c070f536a
7 changed files with 194 additions and 62 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

29
matlab/mcp_func.m
View File

@ -1,4 +1,33 @@
function [res,fjac,domer] = mcp_func(x,jacflag)
% function [res,fjac,domer] = mcp_func(x,jacflag)
% wrapper function for mixed complementarity problem when using PATH
%
% INPUTS
% - x [double] N*T array, paths for the endogenous variables (initial guess).
% - jacflag [scalar] indicator whether Jacobian is requested
%
% OUTPUTS
% - res [double] (N*T)*1 array, residuals of the stacked problem
% - fjac [double] (N*T)*(N*T) array, Jacobian of the stacked problem
% - domer [scalar] errorflag that is 1 if solution is not real
% Copyright (C) 2016 Dynare Team
%
% This file is part of Dynare.
%
% Dynare is free software: you can redistribute it and/or modify
% it under the terms of the GNU General Public License as published by
% the Free Software Foundation, either version 3 of the License, or
% (at your option) any later version.
%
% Dynare is distributed in the hope that it will be useful,
% but WITHOUT ANY WARRANTY; without even the implied warranty of
% MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
% GNU General Public License for more details.
%
% You should have received a copy of the GNU General Public License
% along with Dynare. If not, see <http://www.gnu.org/licenses/>.
global mcp_data
if jacflag

120
matlab/perfect-foresight-models/perfect_foresight_mcp_problem.m
View File

@ -3,23 +3,49 @@ function [residuals,JJacobian] = perfect_foresight_mcp_problem(y, dynamic_functi
maximum_lag, T, ny, i_cols, ...
i_cols_J1, i_cols_1, i_cols_T, ...
i_cols_j,nnzJ,eq_index)
% function [residuals,JJacobian] = perfect_foresight_mcp_problem(x, model_dynamic, Y0, YT,exo_simul,
% params, steady_state, maximum_lag, periods, ny, i_cols,i_cols_J1, i_cols_1,
% i_cols_T, i_cols_j, nnzA)
% computes the residuals and th Jacobian matrix
% for a perfect foresight problem over T periods.
% function [residuals,JJacobian] = perfect_foresight_mcp_problem(y, dynamic_function, Y0, YT, ...
% exo_simul, params, steady_state, ...
% maximum_lag, T, ny, i_cols, ...
% i_cols_J1, i_cols_1, i_cols_T, ...
% i_cols_j,nnzJ,eq_index)
% Computes the residuals and the Jacobian matrix for a perfect foresight problem over T periods
% in a mixed complementarity problem context
%
% INPUTS
% ...
% y [double] N*1 array, terminal conditions for the endogenous variables
% dynamic_function [handle] function handle to _dynamic-file
% Y0 [double] N*1 array, initial conditions for the endogenous variables
% YT [double] N*1 array, terminal conditions for the endogenous variables
% exo_simul [double] nperiods*M_.exo_nbr matrix of exogenous variables (in declaration order)
% for all simulation periods
% params [double] nparams*1 array, parameter values
% steady_state [double] endo_nbr*1 vector of steady state values
% maximum_lag [scalar] maximum lag present in the model
% T [scalar] number of simulation periods
% ny [scalar] number of endogenous variables
% i_cols [double] indices of variables appearing in M.lead_lag_incidence
% and that need to be passed to _dynamic-file
% i_cols_J1 [double] indices of contemporaneous and forward looking variables
% appearing in M.lead_lag_incidence
% i_cols_1 [double] indices of contemporaneous and forward looking variables in
% M.lead_lag_incidence in dynamic Jacobian (relevant in first period)
% i_cols_T [double] columns of dynamic Jacobian related to contemporaneous and backward-looking
% variables (relevant in last period)
% i_cols_j [double] indices of variables in M.lead_lag_incidence
% in dynamic Jacobian (relevant in intermediate periods)
% nnzJ [scalar] number of non-zero elements in Jacobian
% eq_index [double] N*1 array, index vector describing residual mapping resulting
% from complementarity setup
% OUTPUTS
% ...
% residuals [double] (N*T)*1 array, residuals of the stacked problem
% JJacobian [double] (N*T)*(N*T) array, Jacobian of the stacked problem
% ALGORITHM
% ...
% None
%
% SPECIAL REQUIREMENTS
% None.
% Copyright (C) 1996-2015 Dynare Team
% Copyright (C) 1996-2016 Dynare Team
%
% This file is part of Dynare.
%
@ -37,46 +63,46 @@ function [residuals,JJacobian] = perfect_foresight_mcp_problem(y, dynamic_functi
% along with Dynare. If not, see <http://www.gnu.org/licenses/>.
YY = [Y0; y; YT];
residuals = zeros(T*ny,1);
if nargout == 2
iJacobian = cell(T,1);
end
YY = [Y0; y; YT];
i_rows = 1:ny;
offset = 0;
i_cols_J = i_cols;
residuals = zeros(T*ny,1);
if nargout == 2
iJacobian = cell(T,1);
end
for it = 2:(T+1)
if nargout == 1
res = dynamic_function(YY(i_cols),exo_simul, params, ...
steady_state,it);
residuals(i_rows) = res(eq_index);
elseif nargout == 2
[res,jacobian] = dynamic_function(YY(i_cols),exo_simul, params, ...
steady_state,it);
residuals(i_rows) = res(eq_index);
if it == 2
[rows,cols,vals] = find(jacobian(eq_index,i_cols_1));
iJacobian{1} = [offset+rows, i_cols_J1(cols), vals];
elseif it == T + 1
[rows,cols,vals] = find(jacobian(eq_index,i_cols_T));
iJacobian{T} = [offset+rows, i_cols_J(i_cols_T(cols)), vals];
else
[rows,cols,vals] = find(jacobian(eq_index,i_cols_j));
iJacobian{it-1} = [offset+rows, i_cols_J(cols), vals];
i_cols_J = i_cols_J + ny;
end
offset = offset + ny;
i_rows = 1:ny;
offset = 0;
i_cols_J = i_cols;
for it = 2:(T+1)
if nargout == 1
res = dynamic_function(YY(i_cols),exo_simul, params, ...
steady_state,it);
residuals(i_rows) = res(eq_index);
elseif nargout == 2
[res,jacobian] = dynamic_function(YY(i_cols),exo_simul, params, ...
steady_state,it);
residuals(i_rows) = res(eq_index);
if it == 2
[rows,cols,vals] = find(jacobian(eq_index,i_cols_1));
iJacobian{1} = [offset+rows, i_cols_J1(cols), vals];
elseif it == T + 1
[rows,cols,vals] = find(jacobian(eq_index,i_cols_T));
iJacobian{T} = [offset+rows, i_cols_J(i_cols_T(cols)), vals];
else
[rows,cols,vals] = find(jacobian(eq_index,i_cols_j));
iJacobian{it-1} = [offset+rows, i_cols_J(cols), vals];
i_cols_J = i_cols_J + ny;
end
i_rows = i_rows + ny;
i_cols = i_cols + ny;
offset = offset + ny;
end
i_rows = i_rows + ny;
i_cols = i_cols + ny;
end
if nargout == 2
iJacobian = cat(1,iJacobian{:});
JJacobian = sparse(iJacobian(:,1),iJacobian(:,2),iJacobian(:,3),T* ...
ny,T*ny);
end
if nargout == 2
iJacobian = cat(1,iJacobian{:});
JJacobian = sparse(iJacobian(:,1),iJacobian(:,2),iJacobian(:,3),T* ...
ny,T*ny);
end

41
matlab/perfect-foresight-models/perfect_foresight_problem.m
View File

@ -3,23 +3,46 @@ function [residuals,JJacobian] = perfect_foresight_problem(y, dynamic_function,
maximum_lag, T, ny, i_cols, ...
i_cols_J1, i_cols_1, i_cols_T, ...
i_cols_j,nnzJ)
% function [residuals,JJacobian] = perfect_foresight_problem(x, model_dynamic, Y0, YT,exo_simul,
% params, steady_state, maximum_lag, periods, ny, i_cols,i_cols_J1, i_cols_1,
% i_cols_T, i_cols_j, nnzA)
% computes the residuals and th Jacobian matrix
% for a perfect foresight problem over T periods.
% function [residuals,JJacobian] = perfect_foresight_problem(y, dynamic_function, Y0, YT, ...
% exo_simul, params, steady_state, ...
% maximum_lag, T, ny, i_cols, ...
% i_cols_J1, i_cols_1, i_cols_T, ...
% i_cols_j,nnzJ)
% computes the residuals and the Jacobian matrix for a perfect foresight problem over T periods.
%
% INPUTS
% ...
% y [double] N*1 array, terminal conditions for the endogenous variables
% dynamic_function [handle] function handle to _dynamic-file
% Y0 [double] N*1 array, initial conditions for the endogenous variables
% YT [double] N*1 array, terminal conditions for the endogenous variables
% exo_simul [double] nperiods*M_.exo_nbr matrix of exogenous variables (in declaration order)
% for all simulation periods
% params [double] nparams*1 array, parameter values
% steady_state [double] endo_nbr*1 vector of steady state values
% maximum_lag [scalar] maximum lag present in the model
% T [scalar] number of simulation periods
% ny [scalar] number of endogenous variables
% i_cols [double] indices of variables appearing in M.lead_lag_incidence
% and that need to be passed to _dynamic-file
% i_cols_J1 [double] indices of contemporaneous and forward looking variables
% appearing in M.lead_lag_incidence
% i_cols_1 [double] indices of contemporaneous and forward looking variables in
% M.lead_lag_incidence in dynamic Jacobian (relevant in first period)
% i_cols_T [double] columns of dynamic Jacobian related to contemporaneous and backward-looking
% variables (relevant in last period)
% i_cols_j [double] indices of variables in M.lead_lag_incidence
% in dynamic Jacobian (relevant in intermediate periods)
% nnzJ [scalar] number of non-zero elements in Jacobian
% OUTPUTS
% ...
% residuals [double] (N*T)*1 array, residuals of the stacked problem
% JJacobian [double] (N*T)*(N*T) array, Jacobian of the stacked problem
% ALGORITHM
% ...
% None
%
% SPECIAL REQUIREMENTS
% None.
% Copyright (C) 1996-2015 Dynare Team
% Copyright (C) 1996-2016 Dynare Team
%
% This file is part of Dynare.
%

10
matlab/perfect-foresight-models/perfect_foresight_solver_core.m
View File

@ -1,5 +1,15 @@
function [oo_, maxerror] = perfect_foresight_solver_core(M_, options_, oo_)
%function [oo_, maxerror] = perfect_foresight_solver_core(M_, options_, oo_)
% Core function calling solvers for perfect foresight model
%
% INPUTS
% - M_ [struct] contains a description of the model.
% - options_ [struct] contains various options.
% - oo_ [struct] contains results
%
% OUTPUTS
% - oo_ [struct] contains results
% - maxerror [double] contains the maximum absolute error
% Copyright (C) 2015-2016 Dynare Team
%

30
matlab/perfect-foresight-models/private/initialize_stacked_problem.m
View File

@ -1,7 +1,33 @@
function [options, y0, yT, z, i_cols, i_cols_J1, i_cols_T, i_cols_j, i_cols_1, ...
dynamicmodel] = initialize_stack_solve_algo_7(endogenousvariables, options, M, steadystate_y)
dynamicmodel] = initialize_stacked_problem(endogenousvariables, options, M, steadystate_y)
% function [options, y0, yT, z, i_cols, i_cols_J1, i_cols_T, i_cols_j, i_cols_1, ...
% dynamicmodel] = initialize_stacked_problem(endogenousvariables, options, M, steadystate_y)
% Sets up the stacked perfect foresight problem for use with dynare_solve.m
%
% INPUTS
% - endogenousvariables [double] N*T array, paths for the endogenous variables (initial guess).
% - options [struct] contains various options.
% - M [struct] contains a description of the model.
% - steadystate_y [double] N*1 array, steady state for the endogenous variables.
% OUTPUTS
% - options [struct] contains various options.
% - y0 [double] N*1 array, initial conditions for the endogenous variables
% - yT [double] N*1 array, terminal conditions for the endogenous variables
% - z [double] T*M array, paths for the exogenous variables.
% - i_cols [double] indices of variables appearing in M.lead_lag_incidence
% and that need to be passed to _dynamic-file
% - i_cols_J1 [double] indices of contemporaneous and forward looking variables
% appearing in M.lead_lag_incidence
% - i_cols_T [double] columns of dynamic Jacobian related to
% contemporaneous and backward-looking
% variables (relevant in last period)
% - i_cols_j [double] indices of variables in M.lead_lag_incidence
% in dynamic Jacobian (relevant in intermediate periods)
% - i_cols_1 [double] indices of contemporaneous and forward looking variables in
% M.lead_lag_incidence in dynamic Jacobian (relevant in first period)
% - dynamicmodel [handle] function handle to _dynamic-file
% Copyright (C) 2015 Dynare Team
% Copyright (C) 2015-16 Dynare Team
%
% This file is part of Dynare.
%

9
matlab/perfect-foresight-models/sim1.m
View File

@ -3,9 +3,14 @@ function [endogenousvariables, info] = sim1(endogenousvariables, exogenousvariab
% Performs deterministic simulations with lead or lag on one period. Uses sparse matrices.
%
% INPUTS
% ...
% - endogenousvariables [double] N*T array, paths for the endogenous variables (initial guess).
% - exogenousvariables [double] T*M array, paths for the exogenous variables.
% - steadystate [double] N*1 array, steady state for the endogenous variables.
% - M [struct] contains a description of the model.
% - options [struct] contains various options.
% OUTPUTS
% ...
% - endogenousvariables [double] N*T array, paths for the endogenous variables (solution of the perfect foresight model).
% - info [struct] contains informations about the results.
% ALGORITHM
% ...
%

17
matlab/perfect-foresight-models/solve_stacked_problem.m
View File

@ -1,6 +1,19 @@
function [endogenousvariables, info] = solve_stacked_problem(endogenousvariables, exogenousvariables, steadystate, M, options);
% [endogenousvariables, info] = solve_stacked_problem(endogenousvariables, exogenousvariables, steadystate, M, options);
% Solves the perfect foresight model using dynare_solve
%
% INPUTS
% - endogenousvariables [double] N*T array, paths for the endogenous variables (initial guess).
% - exogenousvariables [double] T*M array, paths for the exogenous variables.
% - steadystate [double] N*1 array, steady state for the endogenous variables.
% - M [struct] contains a description of the model.
% - options [struct] contains various options.
%
% OUTPUTS
% - endogenousvariables [double] N*T array, paths for the endogenous variables (solution of the perfect foresight model).
% - info [struct] contains informations about the results.
% Copyright (C) 2015 Dynare Team
% Copyright (C) 2015-16 Dynare Team
%
% This file is part of Dynare.
%
@ -20,7 +33,7 @@ function [endogenousvariables, info] = solve_stacked_problem(endogenousvariables
[options, y0, yT, z, i_cols, i_cols_J1, i_cols_T, i_cols_j, i_cols_1, dynamicmodel] = ...
initialize_stacked_problem(endogenousvariables, options, M, steadystate);
if (options.solve_algo == 10 || options.solve_algo == 11)
if (options.solve_algo == 10 || options.solve_algo == 11)% mixed complementarity problem
[lb,ub,eq_index] = get_complementarity_conditions(M,options.ramsey_policy);
if options.linear_approximation
lb = lb - steadystate_y;