AN INFORMATION-THEORETIC APPROACH TO DETERMINE IMPORTANT ATTRIBUTES FOR ENGINEERING CHANGE EVALUATION

摘要

A proposed hngineering Change (EC) can aliect several lifecycle-wide components. Enterprises plan detailed evaluation of only those changes or effects that may have a significant impact. Using past knowledge of engineering changes can prove effective in determining whether an effect is significant. However, comparing HCs for every single attribute is a combinatorial problem, because an HC is composed of a large number of disparate and interdependent attributes. In this paper, we propose an information-theoretic approach to determine important attributes that should be compared to retrieve past HCs similar to the proposed change. Our approach accounts for unknown interdepen-dencies between attributes of disparate datatypes. We believe that the method discussed in this paper will improve overall efficacy of the Engineering Change Management (HCM) process by facilitating fast and reliable evaluation of a proposed EC.