Preference incorporation in MOEA/D using an outranking approach with imprecise model parameters

Multi-objective Optimization Evolutionary Algorithms (MOEAs) face numerous challenges when they are used to solve Many-objective Problems (MaOPs). Decomposition-based strategies, such as MOEA/D, divide an MaOP into multiple single-optimization sub-problems, achieving better diversity and a approximation of the Pareto front, dealing with some MaOPs. However, these approaches still require one multi-criteria selection problem that will allow Decision-Maker (DM) choose final solution. Incorporating preferences may provide results closer region interest DM. Most proposals integrate in decomposition-based MOEAs prefer progressive articulation over “a priori” incorporation preferences. Progressive methods can hardly work without comparable transitive preferences, significantly increase cognitive effort required On other hand, strategies do not demand judgements from DM but direct parameter elicitation usually is subject imprecision. Outranking have properties them suitably handle non-transitive veto conditions, incomparability, which typical characteristics many real DMs. This paper explores how incorporate MOEA/D using based on interval outranking relations, imprecision preference parameters elicited. Several experiments make it possible analyze proposal's performance benchmark problems compare classic recent, state-of-the-art preference-based decomposition algorithm. In instances, our Region Interest, particularly number objectives increases.