Architecture Evaluation

The architecture evaluator takes a given architecture instance and calculates the defined performance metrics for it. It should run without user interaction.

ADORE provides the following interfaces to connect to evaluation code:

The following guideline can help select the interface appropriate for your SAO problem:

1. If you have to use an external platform, for example a Multidisciplinary Design Analysis & Optimization (MDAO) platform, use a file-based interface.
   • If your MDAO workflow is using CPACS as the central data format, use the CPACS interface.
2. Otherwise, use a Python-based interface.
3. The CFE and NFE are preferred over direct data access, as it:
   • Provides better interoperability between the ADORE model and the evaluation code.
   • It requires less custom coding to parse/convert the data format.
   • It reduces the dependence on ADORE's internal data model.

For all the details and examples regarding evaluation interfaces, including examples and exercises, refer to the following documentation:

and the accompanying: ADORE Evaluation Docs

API references:

• Evaluation and Optimization API
• Architecture data model
• Class Factory Evaluator (CFE)
• XML Node Factory Evaluator (NFE)
• CPACS