In this paper, we consider the problem of automatically assessing sofware quality. We show that we can look at this problem, called Software Quality Assessment (SQA), as a multicriteria decision-making problem. Indeed, just like software is assessed along different criteria, Multi-Criteria Decision Making (MCDM) is about decisions that are based on several criteria that are usually conflicting and non-homogenously satisfied. Nonadditive (fuzzy) measures along with the Choquet integral can be used to model and aggregate the levels of satisfaction of these criteria by considering their relationships. However, in practice, fuzzy measures are difficult to identify. An automated process is necessary and possible when sample data is available. Several optimization approaches have been proposed to extract fuzzy measures from sample data; e.g., genetic algorithms, gradient descent algorithms, and the Bees algorithm, all local search techniques. In this article, we propose a hybrid approach, combining the Bees algorithm and an interval constraint solver, resulting in a focused search expected to be less prone to falling into local results. Our approach, when tested on SQA decision data, shows promise and compares well to previous approaches to SQA that were using machine learning techniques.