Model builders

A model builder defines how a reduced-order model (ROM) is constructed from available full-order model (FOM) samples.

For example, consider an iterative training workflow such as a greedy algorithm. The workflow runs FOM simulations, uses the resulting samples to build a ROM, and evaluates its accuracy. Additional FOM samples are then generated and incorporated until the ROM reaches the desired accuracy.

In this setting, the model_builder interface defines how the ROM is constructed from the collected FOM samples.

Contents:

• Basic model builder tutorial