Chapter 8 Data Structures
8.1 FF_SAVEBORR_GRID Example for Generating Asset Grid
Go back to fan’s MEconTools Toolbox (bookdown), Matlab Code Examples Repository (bookdown), or Math for Econ with Matlab Repository (bookdown).
Examples](https://fanwangecon.github.io/M4Econ/), or** Dynamic Asset This is the example vignette for function: ff_saveborr_grid from the MEconTools Package. This function generates variously spaced savings/borrowing states/choices grid.
8.1.1 Test FF_SAVEBORR_GRID Defaults
Call the function with defaults.
ff_saveborr_grid();
----------------------------------------
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
CONTAINER NAME: mp_container_map ND Array (Matrix etc)
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
i idx ndim numel rowN colN sum mean std coefvari min max
_ ___ ____ _____ ____ ____ ______ ______ ______ ________ ___ ___
ar_fl_saveborr 1 1 2 25 25 1 385.93 15.437 15.324 0.99265 1 50
xxx TABLE:ar_fl_saveborr xxxxxxxxxxxxxxxxxx
c1
______
r1 1
r2 1.0174
r3 1.0982
r4 1.2707
r5 1.5557
r6 1.9707
r7 2.5312
r8 3.2512
r9 4.1434
r10 5.2196
r11 6.4912
r12 7.9687
r13 9.6621
r14 11.581
r15 13.735
r16 16.132
r17 18.781
r18 21.691
r19 24.87
r20 28.324
r21 32.063
r22 36.093
r23 40.421
r24 45.054
r25 50
----------------------------------------
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
CONTAINER NAME: mp_container_map Scalars
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
i idx value
_ ___ _____
grid_evenlog_threshold 1 2 1
grid_log10space_x1 2 3 0.3
grid_log10space_x2 3 4 3
grid_powerspace_power 4 5 2.5 8.1.2 Test FF_SAVEBORR_GRID Default Linear Grid, Log Grid, Power Grid, Threshold Grid
Call the function with defaults.
% Same min and max and grid points
[fl_a_min, fl_a_max, it_a_points] = deal(0,50,25);
% Four types of grid points
st_grid_type = 'grid_linspace';
[ar_fl_saveborr_linspace] = ff_saveborr_grid(fl_a_min, fl_a_max, it_a_points, st_grid_type);
st_grid_type = 'grid_log10space';
[ar_fl_saveborr_log10space] = ff_saveborr_grid(fl_a_min, fl_a_max, it_a_points, st_grid_type);
st_grid_type = 'grid_powerspace';
[ar_fl_saveborr_powerspace] = ff_saveborr_grid(fl_a_min, fl_a_max, it_a_points, st_grid_type);
st_grid_type = 'grid_evenlog';
[ar_fl_saveborr_evenlog] = ff_saveborr_grid(fl_a_min, fl_a_max, it_a_points, st_grid_type);
% draw four types of lines jointly
mt_value = [ar_fl_saveborr_linspace'; ar_fl_saveborr_log10space'; ...
ar_fl_saveborr_powerspace'; ar_fl_saveborr_evenlog'];
ar_row_grid = ["grid linspace", "grid log10space", "grid powerspace", "grid evenlog"];
ar_col_grid = 1:it_a_points;
mp_support_graph = containers.Map('KeyType', 'char', 'ValueType', 'any');
mp_support_graph('cl_st_graph_title') = {'Four Asset Grids with Default Parameters'};
mp_support_graph('cl_st_ytitle') = {'Asset Grid Points'};
mp_support_graph('cl_st_xtitle') = {'Asset Grid Counter'};
mp_support_graph('bl_graph_logy') = true; % do not log
ff_graph_grid(mt_value, ar_row_grid, ar_col_grid, mp_support_graph);
8.1.3 Test FF_SAVEBORR_GRID Log Grid Changing Parameters
Log grid, same min and max, change log X1 and X2 points
% Same min and max and grid points
[fl_a_min, fl_a_max, it_a_points] = deal(0,50,25);
st_grid_type = 'grid_log10space';
% Four types of grid points
mp_grid_control = containers.Map('KeyType','char', 'ValueType','any');
mp_grid_control('grid_log10space_x1') = 0.1;
mp_grid_control('grid_log10space_x2') = 1;
[ar_fl_log10space_a] = ff_saveborr_grid(fl_a_min, fl_a_max, it_a_points, st_grid_type, mp_grid_control);
mp_grid_control('grid_log10space_x1') = 0.1/2;
mp_grid_control('grid_log10space_x2') = 1*2;
[ar_fl_log10space_b] = ff_saveborr_grid(fl_a_min, fl_a_max, it_a_points, st_grid_type, mp_grid_control);
mp_grid_control('grid_log10space_x1') = 0.1/4;
mp_grid_control('grid_log10space_x2') = 1*4;
[ar_fl_log10space_c] = ff_saveborr_grid(fl_a_min, fl_a_max, it_a_points, st_grid_type, mp_grid_control);
mp_grid_control('grid_log10space_x1') = 0.1/6;
mp_grid_control('grid_log10space_x2') = 1*6;
[ar_fl_log10space_d] = ff_saveborr_grid(fl_a_min, fl_a_max, it_a_points, st_grid_type, mp_grid_control);
% draw four types of lines jointly
mt_value = [ar_fl_log10space_a'; ar_fl_log10space_b'; ...
ar_fl_log10space_c'; ar_fl_log10space_d'];
ar_row_grid = [...
"log10space:x1=0.1/1, x2=1", ...
"log10space:x1=0.1/2, x2=2", ...
"log10space:x1=0.1/4, x2=3", ...
"log10space:x1=0.1/6, x2=4"];
ar_col_grid = 1:it_a_points;
mp_support_graph = containers.Map('KeyType', 'char', 'ValueType', 'any');
mp_support_graph('cl_st_graph_title') = {'Asset Grids with Log 10 Grid Varying Controls'};
mp_support_graph('cl_st_ytitle') = {'Asset Grid Points'};
mp_support_graph('cl_st_xtitle') = {'Asset Grid Counter'};
mp_support_graph('bl_graph_logy') = true; % do not log
ff_graph_grid(mt_value, ar_row_grid, ar_col_grid, mp_support_graph);
8.1.4 Test FF_SAVEBORR_GRID Power Grid Changing Parameters
Log grid, same min and max, change log X1 and X2 points
% Same min and max and grid points
[fl_a_min, fl_a_max, it_a_points] = deal(0,50,25);
st_grid_type = 'grid_powerspace';
% Four types of grid points
mp_grid_control = containers.Map('KeyType','char', 'ValueType','any');
mp_grid_control('grid_powerspace_power') = 1;
[ar_fl_powerspace_a] = ff_saveborr_grid(fl_a_min, fl_a_max, it_a_points, st_grid_type, mp_grid_control);
mp_grid_control('grid_powerspace_power') = 2;
[ar_fl_powerspace_b] = ff_saveborr_grid(fl_a_min, fl_a_max, it_a_points, st_grid_type, mp_grid_control);
mp_grid_control('grid_powerspace_power') = 4;
[ar_fl_powerspace_c] = ff_saveborr_grid(fl_a_min, fl_a_max, it_a_points, st_grid_type, mp_grid_control);
mp_grid_control('grid_powerspace_power') = 6;
[ar_fl_powerspace_d] = ff_saveborr_grid(fl_a_min, fl_a_max, it_a_points, st_grid_type, mp_grid_control);
% draw four types of lines jointly
mt_value = [ar_fl_powerspace_a'; ar_fl_powerspace_b'; ...
ar_fl_powerspace_c'; ar_fl_powerspace_d'];
ar_row_grid = [...
"powerspace:power=1", ...
"powerspace:power=2", ...
"powerspace:power=4", ...
"powerspace:power=6"];
ar_col_grid = 1:it_a_points;
mp_support_graph = containers.Map('KeyType', 'char', 'ValueType', 'any');
mp_support_graph('cl_st_graph_title') = {'Asset Grids with Power Grid Varying Controls'};
mp_support_graph('cl_st_ytitle') = {'Asset Grid Points'};
mp_support_graph('cl_st_xtitle') = {'Asset Grid Counter'};
mp_support_graph('bl_graph_logy') = true; % do not log
ff_graph_grid(mt_value, ar_row_grid, ar_col_grid, mp_support_graph);
8.1.5 Test FF_SAVEBORR_GRID Threshold Grid Changing Parameters
Threshold Grid, Changing Threshold Levels. Initial segments below threshold are linspace, then logspace.
% Same min and max and grid points
[fl_a_min, fl_a_max, it_a_points] = deal(0,50,25);
st_grid_type = 'grid_evenlog';
% Four types of grid points
mp_grid_control = containers.Map('KeyType','char', 'ValueType','any');
mp_grid_control('grid_evenlog_threshold') = 0.50;
[ar_fl_evenlog_a] = ff_saveborr_grid(fl_a_min, fl_a_max, it_a_points, st_grid_type, mp_grid_control);
mp_grid_control('grid_evenlog_threshold') = 1.00;
[ar_fl_evenlog_b] = ff_saveborr_grid(fl_a_min, fl_a_max, it_a_points, st_grid_type, mp_grid_control);
mp_grid_control('grid_evenlog_threshold') = 2;
[ar_fl_evenlog_c] = ff_saveborr_grid(fl_a_min, fl_a_max, it_a_points, st_grid_type, mp_grid_control);
mp_grid_control('grid_evenlog_threshold') = 5;
[ar_fl_evenlog_d] = ff_saveborr_grid(fl_a_min, fl_a_max, it_a_points, st_grid_type, mp_grid_control);
% draw four types of lines jointly
mt_value = [ar_fl_evenlog_a'; ar_fl_evenlog_b'; ...
ar_fl_evenlog_c'; ar_fl_evenlog_d'];
ar_row_grid = [...
"evenlog:threshold=0.5", ...
"evenlog:threshold=1.0", ...
"evenlog:threshold=2.0", ...
"evenlog:threshold=5.0"];
ar_col_grid = 1:it_a_points;
mp_support_graph = containers.Map('KeyType', 'char', 'ValueType', 'any');
mp_support_graph('cl_st_graph_title') = {'Asset Grids with Threshold Grid Varying Controls'};
mp_support_graph('cl_st_ytitle') = {'Asset Grid Points'};
mp_support_graph('cl_st_xtitle') = {'Asset Grid Counter'};
mp_support_graph('bl_graph_logy') = true; % do not log
ff_graph_grid(mt_value, ar_row_grid, ar_col_grid, mp_support_graph);
8.2 FX_PERTURB_LOGN Perturb Parameter with Logn Scalar
Go back to fan’s MEconTools Toolbox (bookdown), Matlab Code Examples Repository (bookdown), or Math for Econ with Matlab Repository (bookdown).
Examples](https://fanwangecon.github.io/M4Econ/), or** Dynamic Asset This is the example vignette for function: ff_perturb_logn from the MEconTools Package. This function randomly perturb some existing parameter. See Randomly Perturb Some Parameter Value with Varying Magnitudes.
8.2.1 Test FX_PERTURB_LOGN Defaults
Call the function with defaults.
ff_perturb_logn();
----------------------------------------
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
CONTAINER NAME: mp_container_map Scalars
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
i idx value
_ ___ _______
ar_logn_sd 1 1 0.20534
fl_draw_znorm 2 2 -0.343
logn_coef_of_var 3 3 24.35
param_original 4 4 5
param_perturbed 5 5 4.9296
scaler_0t1 6 6 0.10088.2.2 Test FX_PERTURB_LOGN with Different Draws and How much to Perturb
Call the function with defaults.
% Collect
mp_container_map = containers.Map('KeyType','char', 'ValueType','any');
% Loop over different scalars
param_original = 5;
ar_scaler_0t1 = linspace(0,1,11);
it_scalar_ctr = 0;
for scaler_0t1=ar_scaler_0t1
it_scalar_ctr = it_scalar_ctr + 1;
% Generate differently perturbed parameters
ar_param_perturbed = NaN(1,5000);
for it_rand_seed=1:5000
param_perturbed = ff_perturb_logn(param_original, it_rand_seed, scaler_0t1);
ar_param_perturbed(it_rand_seed) = param_perturbed;
end
% Collect
mp_container_map(['PERTURB_SCALAR_' num2str(scaler_0t1)]) = ar_param_perturbed;
end
% Display
ff_container_map_display(mp_container_map);
----------------------------------------
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
CONTAINER NAME: mp_container_map ND Array (Matrix etc)
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
i idx ndim numel rowN colN sum mean std coefvari min max
__ ___ ____ _____ ____ ____ ___________ _______ _________ __________ _______ ______
PERTURB_SCALAR_0 1 1 2 5000 1 5000 25000 4.9999 0.0022357 0.00044715 4.9922 5.0095
PERTURB_SCALAR_0_1 2 2 2 5000 1 5000 24963 4.9926 0.20802 0.041665 4.2759 5.8859
PERTURB_SCALAR_0_2 3 3 2 5000 1 5000 24889 4.9779 0.62106 0.12476 2.838 7.6451
PERTURB_SCALAR_0_3 4 4 2 5000 1 5000 24756 4.9512 1.3692 0.27654 0.23359 10.831
PERTURB_SCALAR_0_4 5 5 2 5000 1 5000 24521 4.9041 2.6921 0.54893 -4.3712 16.465
PERTURB_SCALAR_0_5 6 6 2 5000 1 5000 24098 4.8196 5.0651 1.0509 -12.632 26.572
PERTURB_SCALAR_0_6 7 7 2 5000 1 5000 23303 4.6606 9.5301 2.0448 -28.175 45.588
PERTURB_SCALAR_0_7 8 8 2 5000 1 5000 21664 4.3328 18.737 4.3245 -60.225 84.8
PERTURB_SCALAR_0_8 9 9 2 5000 1 5000 17645 3.529 41.309 11.705 -138.8 180.93
PERTURB_SCALAR_0_9 10 10 2 5000 1 5000 3041.2 0.60824 123.33 202.77 -424.33 530.27
PERTURB_SCALAR_1 11 11 2 5000 1 5000 -2.0181e+06 -403.62 11475 -28.431 -39942 488788.3 FF_SAVEBORR_GRID Example for Generating Asset Grid
Go back to fan’s MEconTools Toolbox (bookdown), Matlab Code Examples Repository (bookdown), or Math for Econ with Matlab Repository (bookdown).
This is the example vignette for function: ff_nonimg_posnegbd from the MEconTools Package. This function checks for valid domain for function that generates real-valued outcomes, and identifies values along the domain that generates positive and negative Values.
8.3.1 Test FF_NONIMG_POSNEGBD Defaults
Call the function with defaults.
ff_nonimg_posnegbd();
FF_NONIMG_POSNEGBD;it_exit_condition=6;bl_has_increase=0;bl_has_decrease=1;bl_has_constant=0;bl_has_pos=0;bl_has_neg=1;bl_has_zero=0
8.3.2 Test FF_NONIMG_POSNEGBD with Log(x)
Testing the function with log(x)
% Same min and max and grid points
[fl_x_min, fl_x_max, it_eval_points, it_eval_max_round, bl_loop] = deal(-5, 5, 10, 3, true);
[bl_verbose, bl_timer] = deal(true, true);
fc_eval = @(x) log(x);
% Solve
[ar_x_points_noimg, ar_obj_eval_noimg, aar_obj_eval_noimg] = ...
ff_nonimg_posnegbd(fl_x_min, fl_x_max, fc_eval, it_eval_points, it_eval_max_round, bl_loop, ...
bl_verbose, bl_timer);
FF_NONIMG_POSNEGBD;it_exit_condition=1;bl_has_increase=1;bl_has_decrease=0;bl_has_constant=0;bl_has_pos=1;bl_has_neg=1;bl_has_zero=0