A Novel Change Feature-Based Approach to Predict the Impact of Current Proposed Engineering Change
Jinping Chen,Shusheng Zhang,Mingwei Wang,Changhong Xu
DOI: https://doi.org/10.1016/j.aei.2017.06.002
IF: 8.8
2017-01-01
Advanced Engineering Informatics
Abstract:The past engineering change (EC) knowledge can be reused to evaluate the impact of current proposed EC, which is gradually accepted as an effective strategy for engineers to handle EC businesses in enterprises. However, the existing approaches to evaluate EC impact are still time-consuming and complex. So this paper proposes a novel change feature-based approach to predict the impact of current proposed EC. Firstly, the related concepts of change feature are defined. Secondly, the working flow of proposed approach is introduced. Afterwards, a mathematical model is constructed for the prediction of EC impact. Finally, an application case verifies the feasibility of the proposed approach, and the evaluation against two state-of-the-art approaches (namely Mehta’s approach and k-Nearest Neighbor approach) has been done. The results of evaluation show that our approach is better than the two approaches in terms of three indexes: (a) the success rate of prediction, (b) the time of prediction, and (c) the loss function.