แบบจำลองการจัดลำดับความต้องการก่อนหลังสำหรับความต้องการขนาดใหญ่

Abstract

In a software development project, rearranging order of requirements before moving forward to the implementation phase is a challenging for software developers. Because requirement is the basic material for the rest of the software development process, therefore, if the core requirements are mistakenly identified, the software is developed in the wrong direction. It cannot meet the user expectation. Due to the user satisfaction is a measure of the quality of the software, so the software quality is dropped. In order to maximize the user satisfaction under constrain of budget, time and resources, the developers need to prioritize the requirement to identify the core requirements used in project planning. Although, there are many researchers studied on requirements prioritization, however the large scale of requirement- at least 100 requirements or more is one of the most significant current discussions in software engineering. This thesis investigates on designing a requirement prioritizing model to handle the large scale of requirements. This model aims to minimize effort and maximize the accuracy of priority sequence. There are 4 processes to study, including (1) To analyze and propose model, (2) To elicit a traditional prioritizing method, (3) To compare proposed method with traditional method and (4) To evaluate the result. In the first process, the literature review was conducted, and then designed the model. The proposed model composes of 5 steps. They are Priority grouping, Criteria weighting, Estimate requirement by criteria using Weigers’s method, calculate priority and sorting by Binary Search Tree.

The second process aims to investigate on the traditional method and gathering the requirement list by interviewing the case study. The third process was implemented to compare the proposed model with the traditional method by designed experiment. The experiment is conducted in 2 software development projects. The first project is mobile application development with 118 requirements and the second project is web application development with 102 requirements. The expected result is effort evaluation. In the last process, One week after executed the experiment, the accuracy evaluation will be conducted. This experiment focuses on assessing the accuracy of priority sequence of requirements. All interviewee have to choose the most accurate sequence of requirements without realizing on which sequence is from which method (Blind Test). These findings further support the idea that the proposed model can minimize the consumed effort and maximize accuracy of the requirement priority sequence. Moreover, this integrated model can handle the large number of requirements. Based on experiment result, it is interesting to note that effort in all cases have been reduced more than 50%. Regarding the first project, not only the effort is reduced by 53.6%, 63.33% of interviewees from the first project believe that the proposed model provides more accuracy than the traditional one. According to the second project, 33.33% of interviewees from the second project believe that the proposed model provides more accuracy than the traditional one. Moreover, 59.16% of consumed effort for the second project is reduced dramatically.