« Previous « Start » Next »
7 TOMVIEW Utility Functions
The solvers in the TOMVIEW /SOL (MINOS, NLSSOL, SNOPT and NPSOL)
package makes it possible to check the derivatives provided by the
gradient and Jacobian routines.
The sizes and validity is automatically checked by TOMVIEW.
7.1 tomRun
Purpose
General multi-solver driver routine for TOMVIEW.
Description of Inputs
| |
| Solver |
The name of the solver that should be used to optimize the problem. |
| |
| Prob |
Problem reference, see Table 11. |
| |
Description of Outputs
| Result |
Structure with result from optimization, see Table
17. |
Description
The driver routine
tomRun from the block diagram.
7.2 probInit
Purpose
General routine for creating a problem from the test sets.
Description of Inputs
| |
| Problem Type |
The problems set to create a problem from. |
| |
| Problem Number |
Problem number from the test set. |
| |
Description of Outputs
| Problem |
Reference to the Problem object. |
Description
The
probInit routines is used to access the extensive test
sets that are part of the TOMVIEW Base Module.
7.3 Modify
After initializing a problem with TOMVIEW it is possible to modify
parts of it. There are modify, replace, add and remove functions
that act on the problem data. With the modify functions parts of the
problem input may be adjusted, while the replace functions
completely overwrites existing inputs. The add/remove functions make
it possible to add/remove linear constraints.
7.3.1 modify_b_L
Purpose
Modifies lower bounds for the linear constraints,
b_L.
Description of Inputs
| b_L_part |
A value vector with the new elements in b_L. |
| |
| idx |
An integer vector with the indices in b_L to modify. The indices must be non-negative integers. |
7.3.2 modify_b_U
Purpose
Modifies the upper bound of the linear constraints,
b_U.
Description of Inputs
| b_U_part |
A value vector with the new elements in b_U. |
| |
| idx |
An integer vector with the indices in b_U to modify. The indices must be non-negative integers. |
7.3.3 modify_c_L
Purpose
Modifies lower bounds for the nonlinear constraints,
c_L.
Description of Inputs
| c_L_part |
A value vector with the new elements in c_L. |
| |
| idx |
An integer vector with the indices in c_L to modify. The indices must be non-negative integers. |
7.3.4 modify_c_U
Purpose
Modifies upper bounds for the nonlinear constraints,
c_U.
Description of Inputs
| c_U_part |
A value vector with the new elements in c_U. |
| |
| idx |
An integer vector with the indices in c_U to modify. The indices must be non-negative integers. |
7.3.5 modify_x_0
Purpose
Modifies start values for the decision variables
x,
x_0.
Description of Inputs
| x_0_part |
A value vector with the new elements in x_0. |
| |
| idx |
An integer vector with the indices in x_0 to modify. The indices must be non-negative integers. |
7.3.6 modify_x_L
Purpose
Modifies lower bounds for the decision variables
x,
x_L.
Description of Inputs
| x_L_part |
A value vector with the new elements in x_L. |
| |
| idx |
An integer vector with the indices in x_L to modify. The indices must be non-negative integers. |
7.3.7 modify_x_U
Purpose
Modifies lower bounds for the decision variables
x,
x_U.
Description of Inputs
| x_U_part |
A value vector with the new elements in x_U. |
| |
| idx |
An integer vector with the indices in x_U to modify. The indices must be non-negative integers. |
Purpose
Replaces the linear constraints matrix,
A.
Description of Inputs
| A |
Matrix containing the new linear constraints. |
| |
| b_L |
(optional) Vector containing lower bound of the new linear constraints. |
| |
| b_U |
(optional) Vector containing upper bound of the new linear constraints. |
7.3.9 replace_b_L
Purpose
Replaces the lower bounds for the linear constraints,
b_L.
Description of Inputs
| b_L |
A vector with lower bounds for the linear constraints. |
Purpose
Replaces the upper bounds for the linear constraints,
b_U.
Description of Inputs
| b_U |
A vector with upper bounds for the linear constraints. |
Purpose
Replaces the lower bounds for the nonlinear constraints,
c_L.
Description of Inputs
| c_L |
A vector with lower bounds for the nonlinear constraints. |
Purpose
Replaces the upper bounds for the nonlinear constraints,
c_U.
Description of Inputs
| c_U |
A vector with upper bounds for the nonlinear constraints. |
Purpose
Replaces the start value vector for the decision variables
x,
x_0.
Description of Inputs
| x_0 |
A vector with the new start values. |
Purpose
Replaces the lower bound to the decision variables
x,
x_L.
Description of Inputs
| x_L |
A vector with the lower bounds. |
Purpose
Replaces the upper bound to the decision variables
x,
x_U.
Description of Inputs
| x_U_part |
A vector with the upper bounds. |
Purpose
Adds linear constraints to the problem.
Description of Inputs
| A_add |
A matrix containing the additional linear constraints. |
| |
| b_L_add |
A vector with the additional lower bounds for the linear constraints. If b_L is not defined as input a default vector with values −∞ is used. |
| |
| b_U_add |
A vector with the additional upper bounds for the linear constraints. If b_U is not defined as input a default vector with values ∞ is used |
Purpose
Removes linear constraints from the problem.
Description of Inputs
| idx |
An integer vector with the row indices for the linear constraints to remove. |
7.3.18 change_NumDiff
Purpose
Changes the problem flag Prob.NumDiff. If set to 1 numerical
differentiation will be used for for the gradient
g (done
automatically if not given).
Description of Inputs
| NumDiff |
An integer either 0 or 1. NumDiff=1 for numerical differentiation and NumDiff=0 for default (use analytical g if given). |
See
8 for more documentation.
7.3.19 change_ConsDiff
Purpose
Changes the problem flag Prob.ConsDiff. If set to 1 numerical
differentiation will be used for for the constraint Jacobian
J
(done automatically if not given).
Description of Inputs
| ConsDiff |
An integer either 0 or 1. ConsDiff=1 for numerical differentiation and ConsDiff=0 for default (use analytical dc if given). |
See
8 for more documentation.
7.4 Example of TOMVIEW Modify
The following example illustrates how to modify an already existing
problem in LabVIEW. The example is a continuation of the example
problem in section
4.1.1. For example, one might want to add
upper bounds for the decision variable
x after initializing the
problem. One can achieve this by using the TOMVIEW replace
functions. See the Block Diagram in the following figure.
Figure 9:
Modification of example problem.
In the replace block upper bounds are set for
x. In the next
figure the Front Panel is displayed.
Figure 10:
Modification of example problem, Front Panel.
In the Front Panel the new variable
xU is defined. As shown the
new problem has a different solution compared to the same problem
without the upper bounds
xU (the upper bounds were +∞
before).
In a similar way one can replace already existing variables. For
example the matrix
A or the profit vector
c.
« Previous « Start » Next »
TOMVIEW
