Applying statistical approach to check the consistency of pairwise comparison matrices during software requirements prioritization process

摘要

In the field of software engineering, multi-criteria decision making (MCDM) methods have been applied during software requirements prioritization (SRP) process. Among various MCDM methods, analytic hierarchy process is a desired technique to compute the ranking values of the software requirements (SRs). These ranking values are employed to choose the SRs that would be executed during different releases of the software. During SRP, stakeholders specify their preferences on SRs in a matrix, which is known as pairwise comparison matrix (PCM). This matrix is used to compute the ranking values of the SRs. The ranking values of the SRs would be reliable only when the PCMs are consistent. Based on our literature review, we identify that during SRP process less attention is given to statistical based approaches to examine the consistency of PCM. Therefore, to address this issue we proposed a method for SRP which is pliable and genuine because it sets a relevant significance level according to the size of the PCMs. Finally, the proposed method is discussed by considering the SRs of an institute examination system.