Top-Down Evaluation of Reusable Goal Models

Top-down evaluation of goal models allows designers to find the optimal set of solutions that would satisfy the stakeholders of their system and their goals. However, the computational complexity of top-down evaluation increases with the growing size of goal models and goal model reuse hierarchies, when goal models are used in collaboration with other modeling formalisms that may impose some external constraints on them (e.g., feature models). This paper (i) introduces novel modeling constructs and goal prioritization methods into the existing bottom-up evaluation algorithm for its adaptation to top-down evaluation of goal models, (ii) introduces an algorithm to propagate top-level importance values down the reuse hierarchy to benefit from reuse boundaries and allow the goal model of each reuse level to be evaluated individually for the top-down evaluation in the whole reuse hierarchy (i.e., without having to backtrack through the entire reuse hierarchy), and (iii) shows the feasibility of this novel evaluation via its proof-of-concept implementation in the TouchCORE reuse tool.