| |
|
|
-
More than 100 algorithms for different types of optimization.
-
Fully sparsified code, capable of handling large, sparse problems.
The codes are implemented to
efficiently handle Matlab sparse arrays.
-
Automatic differentiation is integrated in TOMLAB by using the
MAD toolbox. TOMLAB /MAD is an add-on toolbox distributed by Tomlab Optimization.
-
Six numerical differentiation methods, selectable.
The standard difference approach, four different spline
methods (three of these requires the Spline Toolbox), and one approach based on complex variables.
-
Compilable with MathWorks MCC. Standalone and embedded applications are easily created.
-
Tomlab standalone licenses are available for redistribution of
Tomlab in standalone applications.
-
Using the Tomlab graphical user interface (GUI)
all solver parameters and features are selectable.
-
The GUI may be used as a code generator. It creates one m-file and one mat-file with the current settings. It is possible to run and edit the settings further. The generated problem may also be read back to the GUI.
-
The menu system, tomMenu,
is another fast way to set all information,
solve the problem and display information.
-
Advanced use of structures for input and output, making is possible to define the user problem once, and use all suitable solvers suitable to solve the problem.
-
Single problems are solved from the
Matlab command line by using the Tomlab Format in a few lines.
Driver routines tomRun and tomSolve can be used to call any solver.
Example files are included for all areas of optimization.
-
Multiple sets of problems are solved by the Tomlab Init File format.
One initialization file is created, together with files to compute the
functions and constraints. Such sets of problems are easily integrated into the GUI and the menu system.
-
Routines are available that convert the Tomlab Format into the Init
File format.
-
Easy to use Tomlab solvers as black-box routines.
-
Robust solution of ill-conditioned nonlinear least squares problems
with linear constraints using several different solver options.
-
Special treatment of exponential fitting and other types of nonlinear
parameter estimation problems.
Unique algorithms for finding initial values and solving
exponential fitting problems.
-
In two-dimensional subspace, GUI and MENU system plots show iteration steps
and marks for line search trials. Linear and nonlinear constraints are
also plotted.
-
Sampled points in two-dimensional subspace for global optimization are shown,
optionally with a contour plot.
-
Experimental data, initial and final models are plotted in nonlinear
least squares fitting.
-
For more information, see the description of each of the different
products.
|
|
