What's New
View new features for HyperStudy 2025.1.
Altair HyperStudy 2025.1 Release Notes
New Features
- Interact with HyperStudy via Programming Interface
- This release introduces a powerful and easy-to-use Python interface, enabling seamless integration of HyperStudy into workflows and applications.
- Leveraging High Power Computing via Altair One
- A new extension tool that enables submitting HyperStudy jobs from local workstations to Altair One appliances.
- Twin Activate Model
- A new model type to integrate multi-disciplinary, dynamic system models via Altair Twin Activate.
- Auto Feature Reduction in Fit
- HyperStudy FAST algorithm now monitors the sensitivity of each input variable on demand and automatically eliminates them based on the contribution threshold defined by the user.
- Access to Tutorial Files within HyperStudy
- Tutorial tab now allows downloading tutorial files directly within HyperStudy.
Enhancements
- Bounds and Modes tabs are consolidated to Define Input Variables tab.
- Linking files between models via drag-and-drop in Model Resources.
- Defining bounds for multiple Design Variables at once.
Resolved Issues
- There were several graphical issues in Dark Mode.
- Pyfit files in Tutorial HS-4550 were broken.
- XML Data Source tool incorrectly highlighted selected attributes.
- Workbench model failed to import variables and extract results.
Announcements
- HyperStudy classic version is no longer available.
- Twin Activate and romAI integrations are supported on Windows only. However, we are working to enable them for Linux in the future.
Altair HyperStudy 2025 Release Notes
New Features
- Leverage accuracy metrics as responses
- HyperStudy Pyfit model now provides an option to include fit metrics as conventional responses. This allows monitoring accuracy and applying objectives which is useful in fit-based optimization approaches.
- Deploy PhysicsAI model as solver
- A new extension tool to auto-register trained PhysicsAI models as solvers.
Enhancements
- Perform trade-off studies following input variable formats
- Tutorial HS-1100, Setting Up Existing Data Model
- Tutorial HS-1695, Using Hooks in PhysicsAI Model
Resolved Issues
- Using labels instead of model parameters in hst_output.hstp files no longer prevents extracting values of different outputs with same labels.
- Hidden or shown tabs no longer restore to default after reopening the application.
Announcements
- Data Source tool, Hstp Reader, has been deprecated and the updater will convert it to XML reader tool. Please use deprecate flag with 8319 to restore the tool.
- “setDirectory” and “setFileHint” python functions have been removed from External Optimizer API. They have been replaced with getResourcePath(type: RESOURCE_TYPE)-> str on the setup class using the resource types RESOURCE_TYPE_WORK_DIRECTORY and RESOURCE_TYPE_FILE_HINT.