Using Analytic Hierarchy Process (AHP) to Prioritize Software Defects
Richard Scold, Robert H. Laurine · ISAHP proceedings · 1994
This report investigates the application of the Analytic Hierarchy Process (AHP) as a decision support tool to aid software managers in the prioritization of software defects (SODS).The effectiveness of an AHP SODS (prioritization) model for information management systems is discussed and illustrated by means of comparison between priorities set by the manager and those set by the model.Today, software maintenance typically consumes more than half of all software development resources.Effective maintenance management is viewed as an important approach to controlling expanding software systems costs.One aspect of software maintenance management which may be improved upon is the area of software defect prioritization, which is the basis for maintenance planning and maintenance budgets.Effectiveness of a SODS decision support tool was evaluated by means of a correlational study involving an operational information system.First, a sample group of 40 software defects was selected from a target population of 668 using a stratified random sampling technique.Then, two instruments were used to prioritize each sample software defect.The first was an after-the-fact instrument, consisting of the manual process currently used by an organization's software engineers to prioritize software defects.The second instrument was the AHP SODS model.Information system software managers representing domain experts for this study provided the technical and managerial criteria and subjective judgments used in building the AHP SODS model.Early results are encouraging, and indicate that the AHP SODS model is an Accurate and practical decision support tool for prioritization of the more severe (P2 level) software defects.Differences in the prioritization of the lower level (P3 through P5 level) software defects between the AHP SODS model and manual methods remain to be explained by continuing research in the Department of Engineering Management at The George Washington University.