Knowledge structure generation and modularization based on binary matrix factorization in engineering design
- PDF / 2,138,483 Bytes
- 17 Pages / 595.22 x 842 pts (A4) Page_size
- 101 Downloads / 167 Views
DOI 10.1007/s12206-020-1024-4
Journal of Mechanical Science and Technology 34 (11) 2020 Original Article DOI 10.1007/s12206-020-1024-4 Keywords: · Binary matrix factorization · Engineering design · Knowledge structure · Modularization · Text mining
Knowledge structure generation and modularization based on binary matrix factorization in engineering design Shubin Zhao1, Cheng Xu1 and Ruizhe Wang2 1
School of Mechanical Engineering, Nanjing University of Science and Technology, Nanjing, China, Viterbi School of Engineering, Department of Computer Science, University of Southern California, California, USA
2
Correspondence to: Cheng Xu [email protected]
Abstract
Citation: Zhao, S., Xu, C., Wang, R. (2020). Knowledge structure generation and modularization based on binary matrix factorization in engineering design. Journal of Mechanical Science and Technology 34 (11) (2020) 4657~4673. http://doi.org/10.1007/s12206-020-1024-4
Received December 5th, 2019 Revised
May 27th, 2020
Accepted August 12th, 2020
Implementing improved effective knowledge reuse and modular design in the earlier stages of engineering design have been the development trend over several decades. However, knowledge capture in the field either remains heavily human involved or is difficult for forming structured knowledge that can be directly utilized by computers. Moreover, modularization is usually performed when system components are fully understood, which fails to influence the design process in an earlier stage. In this paper, we propose a knowledge representation method to capture the knowledge in the design documents intelligently and organize them in a structured manner. A modularization method using binary matrix factorization is also put forward to optimize the design process. Case studies of the design of electro–hydraulic dram brake and power supply are used to exemplify how modularization can be accomplished by matrix factorization and how knowledge structure is generated with computer support. Finally, several research problems are discussed.
† Recommended by Editor Seungjae Min
1. Introduction
© The Korean Society of Mechanical Engineers and Springer-Verlag GmbH Germany, part of Springer Nature 2020
Engineering design is a complex, knowledge-intensive, and often iterative process that relies heavily on existing design knowledge and experience in the area. The majority of engineering designs are related to modifications of previously proven designs to accommodate new requirements [1]. Therefore, proposing effective ways of processing, representing, and reusing domain knowledge has been highly desirable. While a design process progresses sequentially and can be broadly divided into conceptual design and detailed design stage, it is usually not straightforward, that is, the more refined requirements of the product as the design process moves forward often demonstrate several ill-considered aspects of the earlier design stage. Unsatisfactory design results in numerous process iterations; thus, the goal of effective engine
Data Loading...
