An Integrated Framework for Prioritizing Software Specifications in Requirements Engineering

Abstract

Requirements prioritization is an established activity facilitating good decision makingfor the purpose of coping with complexities that often arise when determining the rightchoice of requirement during requirements engineering process. However, makingprecise and accurate decision for the purpose of averting subsequent system failure is anissue of concern when developing large scale systems. In this paper, we proposed the useof imprecise knowledge-based solutions over precise-knowledge based solutions forprioritizing software requirements to overcome the problem of decision making. In thisregard, our proposed technique is an integration of Fuzzy Multi Criteria Decision Making(FMCDM), similarity measures and target-based approach to requirements prioritizationusing linguistic values of triangular fuzzy numbers. With the integrated framework, thechallenge of making decisions under uncertain conditions are considered. The integratedframework was based on some defined criteria, a three-phased process comprising of fivesteps and a five-point scale used to determine the relative values of requirements. Theresult is the specification of a stepwise process of the computations that can be performedduring decision making by the integrated technique proposed. It is hoped that when thistechnique is implemented, executed, evaluated and validated on case study, a promisingresults will be achieved. For validating the proposed technique, requirements from reallifecase studies can be elicited, analysed, and reconciled for completeness andconsistency.