June 30, 2014
pSeven 3.0 Release
DATADVANCE development team is pleased to announce the release of pSeven 3.0, a new stable version of our platform for automated simulation, process integration, multidisciplinary design optimization, and data analysis.
This release brings many upgrades and new features compared to the previous major release, pSeven 2.4. In version 3.0 you will find redesigned Workspace and Run screens, new workflow configuration options, usability improvements in Edit, enhanced std.Composite blocks finalized std.Kompas3D block, new blocks integrating the OpenTURNS methodology into pSeven, and the updates coming from the latest MACROS version – not to mention a number of smaller features and bugfixes.
One of the most noticeable changes in pSeven 3.0 is a complete redesign of the Workspace and Run screens along with the appearance on the new workflow configuration tool which offers many advanced options and allows creating workflows that are fully usable from the Run screen only.
New Run screen pays more attention to workflow inputs, outputs, and configuration settings.
Workflow run-time settings can now be edited in Run interface which is set up by a workflow designer using the workflow configuration tool in Edit. The configuration specifies the set of parameters – block options and port input values – which are shown in Run and the list of ports that allow monitoring. This way, users no longer need to switch from Run to Edit to change block configuration settings and are not required to dive into workflow details when they just want to change a few options or get more data by monitoring an internal port.
Workflow inputs and outputs are now clearly visible in Run. As before, any input port (block) can be uplinked to a workflow input which allows to set its input value in Run; however, the ability to add a port to configuration parameters now often provides a better solution. As for output ports, the key difference is that workflow output values are now also shown in Run, so main results can be viewed without switching to Analyze. All workflow inputs and outputs are also automatically monitored.
The new workflow configuration tool can simplify complex workflows, hiding the unnecessary detail and as a result allowing users to focus on the real task.
First of all, workflow configuration allows adding any port or option as a run-time parameter. Parameter values are specified in Run and are read when the workflow starts, replacing the settings done in Edit. For a port, the parameter becomes its input value – in most cases, this removes the need to add std.Const blocks. It also makes easier to change such inputs and, most importantly, to set a fixed input value for some block that is a part of a loop (for example, the optimized model in an optimization workflow).
Next, the incomprehensible list of all workflow ports no longer exists in Run. Instead, the ports that allow monitoring are selected in workflow configuration, and only these ports are shown in Run, where you can also switch monitoring for a port on or off with a single click.
Moreover, workflow configuration allows adding aliases and descriptions to parameters (ports and options) and the ports selected for monitoring. Aliases are names shown in Run instead of actual port and option names. This way, workflow designers can help users understand the task-specific meaning of Run parameters without making them study how the selected ports and options are processed in the workflow.
An interesting feature of aliases is that multiple ports or options can be given the same alias, in which case only one control is added to Run and can be used to set multiple options or port values at once.
Finally, the workflow configuration tool can be used in Run to review or change the set of workflow inputs, outputs, and settings.
It also allows to add a workflow description which, in addition to the configuration main tab, is also seen in workflow tooltips in Workspace.
The redesigned Workspace screen now provides quick access to project management functions (such as creating and adding projects, workflows, and reports) and various pSeven guides and examples.
The Edit screen has also received certain usability improvements, including block and link tooltips, annotation function, and better toolbar layout.
Block and link tooltips are shown on hover. They contain information on block inputs and outputs; for links, tooltips provide connection details (linked ports and data flow direction).
Workflow annotations are text comments which can be added right to the workflow graph. They can describe the tasks implemented in a workflow, solution details, or intended usage – much like comments in programming code.
In Analyze, pSeven 3.0 adds 3D surface plots supporting various surface reconstruction methods (2D grid, 2D triangulation, 3D triangulation), gradient surface fill, colorbars and surface transparency.
We have also redesigned plot configuration dialogs so that plots of the same type (for example, all 2D plots or all charts) now use a configuration dialog common for this type.
New dialogs also have better settings layout and overall higher usability – for example, you can now change an existing 2D plot to a line plot with a couple of clicks. Also note that all plots now support fixed axis ranges which can be set in general plot configuration: no more need to manually zoom the plot if you want it to show a selected area only.
The std.Composite block in pSeven 3.0 has been enhanced and now handles such advanced functions as:
- Workflow import from file.
- Run-time data cache.
- Automatic parallel processing.
The workflow import function allows to take any existing workflow and make it a new component of the current workflow, replacing the content of the importing std.Composite block.
Run-time data cache function can store input and output values of an std.Composite block to a file specified and use it to speed up processing. If next input value is found in the cache, the corresponding output value is read from cache too, instead of recalculating it. The cache file is a simple CSV that can be easily used outside pSeven.
Parallel processing function allows selecting an input port which receives a list of values and set up the block so it processes this list in parallel, virtually duplicating the workflow contained in an std.Composite block (creating parallel run-time processes).
pSeven 3.0 also finalizes the development of the std.Kompas3D block and adds completely new blocks that integrate the OpenTURNS methodology – std.Distribution and std.UQ (beta).
The final version of the std.Kompas3D block includes such improvements as:
- The support for KOMPAS-3D documents which contain hierarchy and complex trees.
- Full support for assemblies and units.
- Support for a number of export formats.
- Many bugfixes in document exploration and better error handling.
New blocks std.Distribution and std.UQ allow performing an uncertainty quantification study in a pSeven workflow.
std.Distribution creates probabilistic models to simulate uncertainty of input parameters of the studied model. The block supports both parametric and sample-based distributions.
std.UQ analyzes model output and can be used to obtain outputs distribution or estimate model sensitivity to input uncertainties and its failure probability (output exceeding given threshold). Note that both blocks are currently in beta development stage and are intended for experienced OpenTURNS users.
Finally, since pSeven 3.0 includes the latest release of MACROS, its newest features also become available to pSeven users. In particular:
- Surrogate-based optimization (SBO) method of std.Optimizer now supports mixed-integer problems. To use this feature, at least one of the problem objective or constraint functions should be configured as expensive, in order to switch to SBO.
- std.Optimizer is now able to analyze the behavior of linear and quadratic objectives and constraints and restore analytical functions so they are evaluated internally by the block in order to speed up solving.
- std.ApproxBuilder features new Tensor Gaussian Processes (TGP) approximation technique, which is a further development of the methods first introduced in the Tensor Approximation (TA) technique. TGP modifies the Gaussian Processes (GP) algorithm for big data sets and provides model accuracy evaluation support not available in the original TA or incomplete Tensor Approximation (iTA) techniques.
- std.ApproxBuilder now supports model internal validation for the Mixture of Approximators (MoA) technique.
The updates above are described in more detail in the MACROS 3.0 release announcement.
For a complete list of changes in pSeven 3.0, including minor updates, support information, and details on bugfixes, please see the release changelog in pSeven User Manual or pSeven Help. You can also contact us to receive more information and pSeven updates!