Selection and Prioritization of Software Requirements Applying Verbal Decision Analysis

Abstract

During the software development process, the decision maker (DM) must master many variables inherent in this process. Software releases represent the order in which a set of requirements is implemented and delivered to the customer. Structuring and enumerating a set of releases with prioritized requirements represents a challenging task because the requirements contain their characteristics, such as technical precedence, the cost required for implementation, the importance that one or more customers add to the requirement, among other factors. To facilitate this work of selection and prioritization of releases, the decision maker may adopt some support tools. One field of study already known to solve this type of problem is the Search-Based Software Engineering (SBSE) that uses metaheuristics as a means to find reasonable solutions taking into account a set of well-defined objectives and constraints. In this paper, we seek to increase the possibilities of solving the Next Release Problem using the methods available in Verbal Decision Analysis (VDA). We generate a problem and submit it so that the VDA and SBSE methods try to resolve it. To validate this research, we compared the results obtained through VDA and compared with the SBSE results. We present and discuss the results in the respective sections.

Dealing Hybrid Model Appling in the Selection and Prioritization of Software Requirements Using Verbal Decision Analysis

“…In previous researches [9][10][11][12], we had good results when we compared the results generated by the Zapros III-i method concerning the SBSE methods. In these works, it was reported that the VDA method is no better than the SBSE methods, but the gain of this research is the opening of this new front in the VDA field, leaving this field as an alternative nonexistent to select and prioritize software requirements.…”

“…Therefore, this task in this field is relatively new. The methodology used in this work has already been tested in other scenarios in previous studies [9][10][11][12] and has shown to be efficient to solve qualitatively (VDA) quantitative problems (SBSE) with satisfactory results. For this, it was necessary to create an adaptation methodology for this purpose.…”

“…As stated in the previous section, our methodology classifies requirements into two groups: "implementable" and "non-implementable". As in previous work [12], we invited 12 experts with project management experience to respond to questionnaires regarding the classification of requirements since the SAC method does not have a classification platform. These managers had at least 12 months of experience in the area.…”

“…A more significant number of customers, therefore more restrictions and objectives, significantly increase the difficulty of finding a viable solution to the problem. In a previous work [12], we considered five clients per project; in this work, we will consider seven clients per project. Therefore, we have an increase in complexity.…”

“…This work follows extensive research [9][10][11] that aims to add more characteristics and improvements to a new methodology proposed to support DM. This work proposes the construction of a decision support strategy through the use of a pair of methods (different from those used in previous research how in [12]) in the area of Verbal Decision Analysis (VDA), shown in session 4. In the following sessions, we will detail (SBSE) as the primary strategy to find suitable solutions to this type of problem, and finally, we will present a comparison of results obtained between VDA and SBSE with the appropriate comments and the evolution of this research.…”