Product form feature evolution forecasting based on IGMBPM model. Issue 4 (3rd July 2016)

• Record Type: Journal Article
• Authors: Xu, Qiu Ying; Yang, Ming Lang; Liu, Wei Dong; Liu, Chun Jun; Yan, He Min
• Is Part Of: Computer-aided design and applications. Volume 13:Issue 4(2016)
• Page Start: 431
• Page End: 439
• Publication Date: 2016-07-03
• Subjects: Product form feature -- forecasting -- grey model -- back propagation neural network; Computer-aided design -- Congresses; Computer-aided design -- Periodicals; Engineering design -- Data processing -- Congresses; Engineering design -- Periodicals; 620.00420285
• DOI: 10.1080/16864360.2015.1131531
• Languages: English
• ISSNs: 1686-4360

Abstract

ABSTRACT: In view of the problem of generational product consistent form feature is difficult to be predicted quantitatively, this paper presents a novel approach based on grey theory, Back propagation neural network (BP NN) and Markov chain, which is hereafter called the improved Grey-BP model with Markov chain (IGMBPM model). In the process of forecasting, due that the raw sequence consisted of product form feature points' positions has the characteristics of poor sample, irregular and high volatility, firstly the traditional grey model is improved to be more suitable for the oscillating raw data, and the improved grey model is combined with BP NN for the purpose of enhancing the mutual influence between anterior and posterior data in sequence, in addition Markov chain is used to amend the final prediction results. The radiator grill profile of a certain type of automobile are taken as an example, the results of the IGMBPM model are compared with other models, the former shows better performance, which verifies the effectiveness of the proposed method. Study results can be helpful to both designers and stakeholders working in mature or developing manufacturers. GRAPHICAL ABSTRACT: