Multi-objective Evolutionary Algorithms in Recommender Systems

Recommender systems are beneficial in suggesting items to users by knowing their preferences and, therefore, effectively managing the vast amount of available information. Regarding the classical systems that focus on accuracy, the needs of their users have changed so much that many sometimes-conflicting performance measures now have to be taken into account. Recent research has enhanced the applicability of multi-objective evolutionary algorithms in recommender systems, balancing indicators such as accuracy with other essential ones. This survey provides a listing of recent works that applied MOEAs to the problem of recommender systems and pays special attention to critical areas, such as methodological approaches, goals, datasets, and evaluation strategies. This analysis, beyond the state-of-the-art synthesis, helps in the determination of the problems that are linked to the use of MOEAs and the prospects of the development of future research. The exploration targets aiding progress and innovation in this dynamic field.