open menu

Response Surface Methods in modeFRONTIER

​​A rich set of dedicated algorithms support engineers and researchers in building interpolating or approximating surfaces and predicting the behavior of the whole system for a wide range of operating conditions.

Assessing the response of a complex structure often requires a large number of simulations which can be computationally expensive. The Response Surface Methods in modeFRONTIER generate reliable meta-models able to approximate the multivariate input/output behavior of such multifaceted systems, improving the quality of the design knowledge and accelerating the optimization step based on real physics models.

A rich set of dedicated algorithms support specialists in building interpolating or approximating surfaces and predicting the behavior of the whole system for a wide range of operating conditions.

Within modeFRONTIER the RSM wizard allows training and validation datasets to be quickly built on the basis of existing or new tables. As well, engineers are enabled to import previously created RSMs and to perform a contextual screening task to refine the knowledge about variable correlations before the training of meta-models.

The RSM tool offers in-a-glance insights on the generated model quality and enables the designer to select the portion of designs to be evaluated virtually, for a smart exploitation of available computational resources


Support for Functional Mock-up Interface (FMI)

It is now easy to export RSMs directly from the Design Space as Functional Mock-up Units (FMU), computationally inexpensive models and share them with your team or run them as “blocks” within third-party applications compliant with FMI (eg: Simulink/Matlab, LMS Virtual.Lab AMESim, Simulation X, Dymola, MapleSim, etc.)

Gaussian Processes refactored

Thanks to the complete refactoring of the RSM algorithm Guassian Processes, we ensure now easier configuration, increased prediction accuracy and considerably reduced training time.
Make response surface modelling easier and fast: the new RSM Trainer Node automatically trains multiple models simultaneously and identify the most accurate. READ MORE


 modeFRONTIER sophisticated RSM algorithms lead to
  • the ability to perform hundreds of experiments in seconds
  • superior accuracy of metamodels gained with the Validation tool
  • easy and quick wizard-based set-up
  • single-click switch from virtual to real optimization​

Sign in