OpenMDAO: An Open Source Framework for Multidisciplinary Analysis and Optimization

摘要

This paper describes the progress made in the development of OpenMDAO, an open source framework for performing Multidisciplinary Analysis and Optimization (MDAO). NASA intends to use OpenMDAO to aid in the design of unconventional aircraft, but the general structure and methods may be applied to solve any number of engineering-related design problems. The framework currently supports data passing capabilities, and several example problems have been executed with it. Recent work has focused on enabling the creation of more complex MDAO strategies, such as collaborative optimization and surrogate modeling techniques. An example is presented that demonstrates an implementation of the surrogate model generation using Kriging surrogate models augmented with the expected improvement method.