DADOS: A Cloud-based Data-driven Design Optimization System

Abstract

Abstract This paper presents a cloud-based data-driven design optimization system, named DADOS, to help engineers and researchers improve a design or product easily and efficiently. DADOS has nearly 30 key algorithms, including the design of experiments, surrogate models, model validation and selection, prediction, optimization, and sensitivity analysis. Moreover, it also includes an exclusive ensemble surrogate modeling technique, the extended hybrid adaptive function, which can make use of the advantages of each surrogate and eliminate the effort of selecting the appropriate individual surrogate. To improve ease of use, DADOS provides a user-friendly graphical user interface and employed flow-based programming so that users can conduct design optimization just by dragging, dropping, and connecting algorithm blocks into a workflow instead of writing massive code. In addition, DADOS allows users to visualize the results to gain more insights into the design problems, allows multi-person collaborating on a project at the same time, and supports multi-disciplinary optimization. This paper also details the architecture and the user interface of DADOS. Two examples were employed to demonstrate how to use DADOS to conduct data-driven design optimization. Since DADOS is a cloud-based system, anyone can access DADOS at www.dados.com.cn using their web browser without the need for installation or powerful hardware.

Keywords

• Data-driven
• Optimization
• Cloud-based software
• Design of experiments
• Surrogate model