Knowledge Enhanced Experience Reuse for Requirements Engineering Decision Support (KEEREDECS)

Abstract

Requirements insufficiency is one of the major contributing factors to software projects failure, indicating that Requirement Engineering (RE) process must be applied judiciously to improve requirements quality. Potentially, many activities can be performed during RE process, and each activity can be supported by several techniques. There is general consensus that RE activities and techniques should be customized to suit contextual features like project, process, product etc. Moreover, a holistic view of the process is needed while deciding the RE activities and techniques. Providing automatic support for fulfilling this need is a challenging problem due to non-deterministic and human-oriented nature of decision-making in the domain. The KEEREDECS framework proposed in this work is an automated approach that will makephase wise contributions to RE process. First, a context processing module is used to get project and techniques context parameters values and weights.Second,Fuzzy Case Based Reasoning (FCBR) technique from Artificial Intelligence is used to handle non-deterministic and context-sensitive decision-making with the help of experiential learning. Third, technique set generations module is used to involve the expert judgment to support decision making.The system can run autonomously or with expert involvement. Evaluation ofKEEREDECS indicates that the approach has a good potential to support for informed decision-making leading to better quality of obtained requirements