A Unified Framework for Multidisciplinary Design Optimization Application

Abstract

Research on the development of optimization modeling systems is relatively mature, but for multidisciplinary design optimization (MDO) application there are still some issues of accessibility and high efficiency. Based on the existing MDO algorithms, this paper presents a unified framework for formulating and solving the multidisciplinary optimization problem with various MDO methods in an efficient, simple and friendly manner. A multidisciplinary descriptive language is also proposed at the same time for the unified framework. The framework is designed to make researchers focus on the multidisciplinary optimization problem, rather than on algorithms implementation and software programming. Using the unified framework, researchers can solve the multidisciplinary optimization problem by creating a configuration file without any coding. The framework has integrated a variety of MDO architectures and also provides interface for other developers to add new optimization algorithm. Furthermore, examples demonstrate the usage of the framework as well as the advantages of the unified framework.