New Functionality in Version 4.3a
An Optimization study for gradient-free (derivative-free) optimization can now be added directly from the Study node. The objective function and constraints are added in the study node. The Optimization solver collects the settings necessary for solving optimization problems based on model parameters. This solver allows parameters that control the geometry or mesh sequence, compared to the more limited impact of global constraint variables in the Optimization user interface. For gradient-free optimization, three new optimization methods are available:
- Nelder-Mead (the default)
- Monte Carlo
- Coordinate search
You can specify the objective and the control parameters directly in the Optimization study node’s settings window, and it is also possible to select if the optimization minimizes or maximizes the values of the objective.
New Model in Version 4.3a
- An extension of the Tuning Fork model, the Shape Optimization of a Tuning Fork model (tuning_fork_optimization) shows how to set up an Optimization study to determine the prong length at which the fork vibrates at the standard concert pitch, 440 Hz.