Model-Based Systems Analysis & Engineering: Standard Evaluator

View at publisher
Authors: Joerg M. Gablonsky, Alexander L. Carrere, Ranald M. Engelbeck, Micah Goldade, Jeffery D. Musiak, Eduardo Ocampo, Sean R. Wakayama
Cited by: 0
References: 3

Abstract

This paper introduces the standard-evaluator: an open-source, framework‑independent library that captures, validates, and serializes Multidisciplinary Design, Analysis, and Optimization (MDAO) assemblies, component option metadata, and optimization problem definitions. Implemented with Pydantic schemas and JSON serialized files, and leveraging HDF5 for compact state capture, the standard‑evaluator enables (1) programmatic discovery and machine‑readable description of component initialization options (types, defaults, bounds, units, and provenance), (2) export/import of full assembly interfaces and runtime state independent of any MDAO engine, and (3) textual, version‑controlled definition of optimization studies that can be instantiated into concrete solvers, (4)encapsulation, requiring components, including groups, in a Multidisciplinary Analysis (MDA) to explicitly define inputs and outputs, and defining others as internal to the component. We demonstrate numerous applications using OpenMDAO as a first use-case. The library’s option API reduces human‑error and manual inspection when composing reusable components, simplifies surrogate‑replacement and remote execution workflows by enforcing explicit external interfaces, and facilitates multi‑framework portability to avoid vendor lock‑in. Released under the Apache‑2.0 license, the standard‑evaluator is positioned to accelerate collaborative, reproducible, and traceable MDAO studies across industry and academia, and to serve as a bridge to GUI‑driven assembly tools.

Fields of Study & Subjects

Advanced Multi-Objective Optimization Algorithms Systems Engineering Methodologies and Applications Modeling and Simulation Systems

Publication details

Journal / source
AIAA SciTech Forum
Publication date
8 January 2026