October 10, 2016
pSeven 6.9 Release
DATADVANCE development team is pleased to announce the release of pSeven 6.9, a new version of the design space exploration platform for every expertise.
Major updates in pSeven 6.9 include:
- New Excel integration block enables using Excel calculations in pSeven workflows. Watch demo
- Tiling layout for reports in Analyze replaces old viewer windows with tiles that are snapped to grid, making viewer management much more convenient.
- Functional Mock-up Interface (FMI) support - export pSeven approximation model as a Functional Mock-up Unit for co-simulation.
- Model explorer - a new tool to study behavior of multidimensional models added to the Predictive Modeling Toolkit. Watch demo
- More new features in the Predictive Modeling Toolkit:
- New Split data command to create model training and test data sets faster.
- Extended model predictions dialog enables to generate input samples on the fly when evaluating a model.
- Improved model performance in the Excel DLL export format.
- Shared cache support for Composite blocks - if you use multiple copies of the same Composite block, they can cache inputs and outputs to the same file, sharing results and saving even more CPU cycles.
- Various convenience features:
- The Duplicate command in Edit helps to create block copies. Clone your cached Composite blocks with a single click.
- Edit, Run, and Analyze now keep tab order when you close the project.
- The Save project command to save all workflow and report changes at once.
- Updated examples with the new tiling layout, and new example projects:
- Dimension reduction - the feature extraction method.
- Advanced approximation with data fusion - sample- and blackbox-based.
- Tutorial-style introduction to the Predictive Modeling Toolkit.
- Three-section beam optimization example showing integration with MSC Nastran.
pSeven 6.9 also includes the new version of the pSeven Core library containing the latest updates to modeling and optimization algorithms, in particular:
- New option that allows to specify the methods to use in optimization: configure the problem as single- or multi-objective, enable robust or surrogate based optimization, select global search methods, and other.
- Support for using data samples with non-numeric and missing values as initial data in optimization.
- Improved handling of categorical variables in SmartSelection modeling algorithms.