Window-of-interest based Multi-objective Evolutionary Search for Satisficing Concepts
This work addresses a specific concept-based exploration problem in design optimization, representing an incremental advancement over prior Pareto-based methods.
The paper tackles the problem of identifying design concepts with at least one solution within a pre-defined performance window-of-interest, proposing a WOI-based multi-objective evolutionary algorithm that outperforms sequential exploration in handling numerical difficulties.
The set-based concept approach has been suggested as a means to simultaneously explore different design concepts, which are meaningful sub-sets of the entire set of solutions. Previous efforts concerning the suggested approach focused on either revealing the global front (s-Pareto front), of all the concepts, or on finding the concepts' fronts, within a relaxation zone. In contrast, here the aim is to reveal which of the concepts have at least one solution with a performance vector within a pre-defined window-of-interest (WOI). This paper provides the rational for this new concept-based exploration problem, and suggests a WOI-based rather than Pareto-based multi-objective evolutionary algorithm. The proposed algorithm, which simultaneously explores different concepts, is tested using a recently suggested concept-based benchmarking approach. The numerical study of this paper shows that the algorithm can cope with various numerical difficulties in a simultaneous way, which outperforms a sequential exploration approach.