tbpsg_function_gradient
Contents
tbpsg_function_gradient function calculates PSG function (see PSG Function) gradient on a PSG point.
Syntax
[val] = tbpsg_function_gradient(func_descr, point_name, toolboxstruc_arr)
Description
[val] = tbpsg_function_gradient(func_descr, point_name, toolboxstruc_arr) returns vector with gradient of function described in func_descr on PSG point point_name (see PSG Point in MATLAB) with data packed in structure toolboxstruc_arr (see PSG Data Objects in MATLAB).
Input Arguments
func_descr |
string with description of PSG function (see PSG Function); |
point_name |
name of PSG point (see PSG Point in MATLAB); |
toolboxstruc_arr |
array of PSG data structures (see PSG Data Objects in MATLAB). |
Output Arguments
val |
vector with function gradient. |
Remarks
| • | toolboxstruc_arr should contain all matrices from func_descr as well as point named point_name. |
| • | tbpsg_function_gradient does not calculate gradient of the Linear Multiple function. |
Example
Input data:
H1 = [1 4 8 3;7 -5 4 -6;-2 8 -1 0;0 -3 -4 9];
c1 = [-2;11;6;10];
p1 = [0.2;0.2;0.3;0.3];
a = [1;1;1;1];
alpha = 0.65;
To pack matrix, point and parameter to array of structures "toolboxstruc_arr":
toolboxstruc_arr(1) = tbpsg_matrix_pack('matrix_1', H1, [], c1, p1);
toolboxstruc_arr(2) = tbpsg_point_pack('point_1', a, []);
toolboxstruc_arr(3) = tbpsg_parameter_pack('parameter_alpha', alpha);
Calculate gradient of cvar_risk function with parameter 0.65 at point 'point_1':
val = tbpsg_function_gradient('cvar_risk(parameter_alpha,matrix_1)','point_1',toolboxstruc_arr);
Output:
val =
-4.0000 4.1429 -0.5714 -0.4286
See also
tbpsg_function_value, tbpsg_function_increment,tbpsg_loss_values