An Improved Fuzzy AHP–TOPSIS Framework for Functional Requirements Selection and Rank-Reversal Mitigation

Abstract

Selecting appropriate functional requirements is a critical challenges in software requirements engineering. The difficulty arises from the need to reconcile conflicting stakeholder preferences, handle uncertainty inherent in linguistic assessments, and address the potential for rank reversal during the prioritization process. While existing fuzzy-based methods, especially those relying solely on Fuzzy TOPSIS, can handle vague judgments, they often lead to unstable rankings when new options are introduced or existing ones are removed. This study introduces a refined framework for selecting functional requirements that combines the Extent Fuzzy Analytic Hierarchy Process (FAHP) with the Fuzzy Technique for Order Preference by Similarity to Ideal Solution (TOPSIS). Our framework uses Extent FAHP to determine the relative importance of non-functional requirements through pairwise comparisons in a fuzzy context, while Fuzzy TOPSIS is employed to rank functional requirements based on these calculated weights. We apply this framework to an Institute Examination System case study, evaluating fifteen functional requirements alongside five non-functional ones using triangular fuzzy numbers and linguistic variables. To ensure the framework’s reliability, we validate it through three ranking scenarios: a baseline evaluation, the addition of a new functional requirement, and the removal of an existing one, all aimed at testing ranking stability. The experimental results show that our hybrid approach yields consistent and robust rankings, mitigating rank reversal and addressing uncertainties in stakeholder judgments. When compared to the traditional Fuzzy TOPSIS-based methods, our framework offers more dependable prioritization of requirements, and enhanced decision-making support for software engineers.