Evaluation of Goal Models in Reuse Hierarchies with Delayed Decisions

摘要

Trade-off analysis through goal model evaluation has been a valuable tool for requirements elicitation and analysis. This is also true in the context of reuse. When goal models are used to describe reusable artifacts and to represent the impacts of reusable artifacts on high-level goals and qualities, they can guide the selection of reusable artifacts to build reuse hierarchies. In previous work, we introduced the use of relative contribution values for reusable goal models, while considering constraints imposed by other modeling notations. In this paper, we expand the result of goal model evaluation from the typical single satisfaction value to a range of values that are still possible based on the current task selections. In the context of reuse hierarchies, we call the remaining task selections delayed decisions because they are postponed to a higher level in the reuse hierarchy when more is known about the system under development. The extended algorithm takes into account the delayed decisions and evaluates the best and worst possible results that can be obtained with the task selections that have been made in the entire reuse hierarchy. The distinct levels in the reuse hierarchy are leveraged to manage the computational complexity of this reuse hierarchy-wide evaluation. A proof-of-concept implementation of the novel evaluation algorithm is presented in the concern-oriented software design modeling tool TouchCORE.