The ultimate objective of Zen-Configurator is to find an efficient solution to guide configuration engineers through the complicated and error-prone process such that eventually the overall quality and productivity of the product development can be improved. As the first step towards this goal, we devised a search-based solution, i.e., Zen-DO, for addressing the decision ordering functionality in an efficient and scalable way.
Configuration engineers in this configuration environment are considered as part of the optimization as they receive feedback (on which decisions to make first) from the tool and configuration engineers’ manual decisions trigger Zen-DO to dynamically find other optimized configuration orders based on the remaining variation points to be configured in the whole configuration space (all variation points should be configured and all constraints should be satisfied to various extents). This kind of configuration design is called “User-in-the-loop”.
Based on the results of the empirical study, we suggest using either AVM or (1+1) EA together with our fitness function to solve the decision-ordering problem, since these two algorithms exhibit the best performance as compared to the other algorithms.