A dynamic model for engineering change propagations in multiple product development stages. (16th February 2022)

• Record Type: Journal Article
• Authors: Li, Yulaing; Zhao, Wei; Zhang, Wenqi; Chen, Meng
• Is Part Of: AI EDAM. Volume 36(2022)
• Publication Date: 2022-02-16
• Subjects: Change propagation -- multistage development -- propagation dynamics -- simulation; Engineering design -- Data processing -- Periodicals; Artificial intelligence -- Periodicals; Expert systems (Computer science) -- Periodicals; 620.00420285
• DOI: 10.1017/S089006042100041X
• Languages: English
• ISSNs: 0890-0604

Abstract

Abstract: To accurately predict propagation dynamics for single or multiple change propagations across different product development stages in a sequential or concurrent way is critical for decision-making of implementing change requests. In this paper, a change propagation dynamic model is built based on the compartmentalization of engineering entities into susceptible engineering entities and affected engineering entities (SA), the ordinary differential equations for describing the rate of affected entities with respect to the total ones and the duration for resolving all the changes for every moment are presented by combining the calculations of change impacts with different split and joint junctions. Considering the difficulty of finding analytical solutions to the differential equations, algorithms for sequential and concurrent simulations of change propagations across different development stages, and random and GA (Genetic Algorithm)-based optimal selections of feasible propagation paths are developed to obtain numerical solutions for single and multiple change requests. Simulation results show that change ripples and blossoms can be observed in both sequential and concurrent change propagations, and these propagation patterns are not sensitive to the initial change effect and the threshold value for propagations, while critical change propagation paths and the number of initiated changes have important effects on both concurrent and sequential change propagation process. It is also demonstrated that concurrent propagation strategy is advantageous for processing single or few of initiated changes since it can shorten product redevelopment time, sequential propagation strategy has an advantage of robustness for handling multiple initiated change requests.