Post by dgorissen on Apr 28, 2009 17:20:28 GMT -5
We are happy to announce the 6.1.1 release of the SUrrogate MOdeling
(SUMO) Toolbox.
[/li][li] The SUMO Toolbox is a Matlab toolbox that automatically generates
a surrogate model for a given data source (a simulation code, data set,
Matlab function, ...) within the predefined accuracy, sample budget,
and time limits set by the user.
[/li][li] It will automatically drive your simulation code generating an
approximation model (ANN, SVM, rational function, RBF model, spline,
(Blind-)Kriging, ...) that is as accurate as possible, using as little data
points as possible (since these are usually expensive). Sample
selection is done adaptively (= sequential design)
and the model parameters (e.g., ANN topology, Kriging parameters, ...)
are determined automatically. In this way it is complimentary to the
excellent Surrogates Toolbox developed by Felipe Viana.
[/li][li] While the main focus is on global metamodeling, some optimization support is available in the form of the Efficient Global Optimization (EGO) algorithm with different expected improvement criteria.
[/li][li] An overview presentation is available here: www.sumowiki.intec.ugent.be/images/e/e1/SUMO_presentation.pdf
[/li][li] For more information, screenshots, movies, downloads, etc. see: www.sumowiki.intec.ugent.be/
[/li][li] In this release many important bugs have been fixed and new
features (such as support for multi-objective model generation and
Gaussian Process Models) have been added. Information about the new
features and changes in this release are available here:
www.sumowiki.intec.ugent.be/index.php/Whats_new
www.sumowiki.intec.ugent.be/index.php/Changelog
[/li][li] Questions, problems, ... can be directed to: www.sumowiki.intec.ugent.be/index.php/Contact
Kind regards
The SUMO Lab Team
--
Surrogate Modeling Lab
Ghent University, Belgium
Web: www.sumo.intec.ugent.be
Blog: sumolab.blogspot.com
(SUMO) Toolbox.
[/li][li] The SUMO Toolbox is a Matlab toolbox that automatically generates
a surrogate model for a given data source (a simulation code, data set,
Matlab function, ...) within the predefined accuracy, sample budget,
and time limits set by the user.
[/li][li] It will automatically drive your simulation code generating an
approximation model (ANN, SVM, rational function, RBF model, spline,
(Blind-)Kriging, ...) that is as accurate as possible, using as little data
points as possible (since these are usually expensive). Sample
selection is done adaptively (= sequential design)
and the model parameters (e.g., ANN topology, Kriging parameters, ...)
are determined automatically. In this way it is complimentary to the
excellent Surrogates Toolbox developed by Felipe Viana.
[/li][li] While the main focus is on global metamodeling, some optimization support is available in the form of the Efficient Global Optimization (EGO) algorithm with different expected improvement criteria.
[/li][li] An overview presentation is available here: www.sumowiki.intec.ugent.be/images/e/e1/SUMO_presentation.pdf
[/li][li] For more information, screenshots, movies, downloads, etc. see: www.sumowiki.intec.ugent.be/
[/li][li] In this release many important bugs have been fixed and new
features (such as support for multi-objective model generation and
Gaussian Process Models) have been added. Information about the new
features and changes in this release are available here:
www.sumowiki.intec.ugent.be/index.php/Whats_new
www.sumowiki.intec.ugent.be/index.php/Changelog
[/li][li] Questions, problems, ... can be directed to: www.sumowiki.intec.ugent.be/index.php/Contact
Kind regards
The SUMO Lab Team
--
Surrogate Modeling Lab
Ghent University, Belgium
Web: www.sumo.intec.ugent.be
Blog: sumolab.blogspot.com