Modelo multicritério para priorização de complexidade de requisitos de software

Abstract

The increasing use of information technology in the world scenario, with high investment in the area and with priority for software, means that to develop them, they have an even greater responsibility to deliver a quality product on time and budget. In this context, this work presented an objective to develop a multicriteria model to prioritize software requirements based on its complexity from the technical team's point of view. The proposed model was developed based on a systematic literature review on requirements prioritization and software estimation to obtain the impact factors. These factors were grouped and then evaluated for importance by the experts for the context of software development in general. The AHP multicriteria method was used to generate the weight of each impact factor evaluated. The Fuzzy TOPSIS method was used to process the qualitative evaluations of the impact factors regarding their complexity with the requirements. A practical application was performed in the project management of web software in a company in the southwestern region of Paraná. The complexity of each impact factor was evaluated for each of the ten requirements of the project by three members with different specialty profiles. The Fuzzy TOPSIS multicriteria method was used to generate the ordering of the requirements by the complexity evaluated. The requirement 5 RQ5) (0.631) ranked first, followed by RQ8 (0.622) in second and RQ9 (0.609) in third, RQ7 (0.602) in fourth, RQ1 (0.581) in fifth. Then RQ2 (0, 499), RQ10 (0.478), RQ6 (0.104), and RQ3 (0.101) in the sixth, seventh, eighth, and ninth positions, respectively, thus RQ4 (0.051) was in the last position regarding its complexity with the impact factors. The impact factors that highlighted the complexity of the ten software requirements evaluated were: runtime restriction (FI1), required software reliability (FI5), product complexity (FI7), documentation compatible with life cycle needs (FI9), application experience (FI13), language and tool experience (FI15), use of software tools (FI17) and required development schedule (FI19). It was found through the constructs of the Technology Acceptance Model (TAM) of Davis (1989) that those involved in the application of the model developed to have a good acceptance for perceiving the Perceived Usefulness and Ease of Use, which effectively contributes by the Attitude and Behavioral Intention for the Actual Use. However, it has received suggestions for adjustments regarding the possibility of choice and several factors.