Quick vs Benchmark vs More Precise (Savings Dynamic Programming Problem)

back to Fan's Dynamic Assets Repository Table of Content.

Testing the ff_az_vf_vecsv program for solving the savings only dynamic programming problem.

Computational speed is determined by the number of asset and shock grid points. Here we run the model with quick, benchmark and more precise simulations with increasing grid points for shocks and asset.

  1. quick: fast run
  2. benchmark: default as set in ffs_az_set_default_param
  3. more precise: increase grid count for shocks and asset

@seealso

Contents

Quick Solution

it_param_set = 4;
[param_map, support_map] = ffs_az_set_default_param(it_param_set);

% Simulation Accuracy
param_map('it_a_n') = 100;
param_map('it_z_n') = 11;

% Display Parameters
support_map('bl_display') = false;
support_map('bl_display_final') = false;
support_map('bl_time') = true;
support_map('bl_profile') = false;

% Call Program
ff_az_vf_vecsv(param_map, support_map);

% Snap
snapnow;
close all;
Elapsed time is 0.069095 seconds.

Benchmark Solution

it_param_set = 4;
[param_map, support_map] = ffs_az_set_default_param(it_param_set);

% Simulation Accuracy
param_map('it_a_n') = 750;
param_map('it_z_n') = 15;

% Display Parameters
support_map('bl_display') = false;
support_map('bl_display_final') = false;
support_map('bl_time') = true;
support_map('bl_profile') = false;

% Call Program
ff_az_vf_vecsv(param_map, support_map);

% Snap
snapnow;
close all;
Elapsed time is 1.697590 seconds.

More Precise Solution

it_param_set = 4;
[param_map, support_map] = ffs_az_set_default_param(it_param_set);

% Simulation Accuracy
param_map('it_a_n') = 2250;
param_map('it_z_n') = 27;

% Display Parameters
support_map('bl_display') = false;
support_map('bl_display_final') = false;
support_map('bl_time') = true;
support_map('bl_profile') = false;

% Call Program
ff_az_vf_vecsv(param_map, support_map);

% Snap
snapnow;
close all;
clear all;
Elapsed time is 28.395729 seconds.