A partial order assimilation approach for software requirements prioritization

The goal of any automated requirement prioritization process is to reduce the difference between the true and estimated ranking. It is arguable that to what extent, using stakeholders' feedback will reduce that difference. Moreover, incorporating requirement ranking functions such as value, cost, risk, etc. along with the feedback is another concern. This paper proposes a scheme for requirement prioritization addressing the feedback issue along with a highly significant ranking function that have decisive impact on true ranking. The approach initially creates partial orders on the basis of stakeholders' feedback and then merges those orders using binary search technique. Two variations of proposed method were implemented varying the number of elicited requirement pairs which were ordered based on stakeholders' preference. The variations were analyzed on the basis of displacement for each requirement from true to estimated rank. Results indicate that for projects having significant ranking functions, decreasing stakeholders' feedbacks could help in reducing the ranking difference.