Requirement-driven engineering change management in product design

Abstract

Product design change is often triggered by external factors. As product is composed of various components and complex relations among them, predicting change propagation of the upstream component and subsequent influence on product design change is a great challenge. In current researches, most of the methods for engineering change (EC) prediction are focused on a single index, e.g., cycle or workload. Besides, logical relations between product components are rarely considered in the product design change. In this paper, an evaluation model of change propagation incorporating multiple indices and logical relations is proposed to manage the ECs. The primary result of this work is the provision of a development support to efficiently acquire the optimal design change solution when a changed requirement is proposed. Requirement is regarded as the source of design change. Requirement-component relation is established, and component-component relation is constructed in terms of interface types between components. Logical relations (i.e., AND, OR, XOR) between downstream components are proposed and the interface types are applied to judge the logical relation. Requirement-component relation, logical relation between components are applied to generate the change propagation paths. Then ant colony optimization (ACO) is used to optimize the generated change propagation paths and the optimal design change solution is demonstrated to the designers automatically. Finally, a prototype software system is developed and a case study is detailed.